You are on page 1of 24

Developmental Science 18:1 (2015), pp 123 DOI: 10.1111/desc.



Development and the epigenome: the synapse of
geneenvironment interplay
W. Thomas Boyce1,3 and Michael S. Kobor2,3
1. Departments of Pediatrics and Psychiatry, University of California, San Francisco, USA
2. Centre for Molecular Medicine and Therapeutics, Department of Medical Genetics, University of British Columbia, Canada
3. Child and Brain Development Program, Canadian Institute for Advanced Research, Canada

This paper argues that there is a revolution afoot in the developmental science of geneenvironment interplay. We summarize, for
an audience of developmental researchers and clinicians, how epigenetic processes chromatin structural modifications that
regulate gene expression without changing DNA sequences may offer a strong, parsimonious account for the convergence of
genetic and contextual variation in the genesis of adaptive and maladaptive development. Epigenetic processes may play a
plausible explanatory role in understanding: divergent trajectories and sexual dimorphisms in brain development; statistical
interactions between genes and environments; the biological embedding of early psychosocial adversities; the linkages of such
adversities to disorders of mental health; the striking individual variation in the strength of those linkages; the molecular origins
of critical and sensitive periods; and the transgenerational inheritance of risk and protection. Taken together, these arguments
converge in a claim that epigenetic processes constitute a promising and illuminating point of connection a synapse between
genes and environments.

Introduction ing of molecular neurotransmission between neurons;

the electrophysiology of circuitry formation and activa-
At the October 1889 meeting of the German Anatomical tion; and the use of psychopharmacological agents to
Society in Berlin, celebrated Spanish neuroanatomist alter functional relations among brain structures.
Santiago Ram on y Cajal presented his histological slides This paper will argue that the emerging science of
of brain tissue stained using the technique developed by epigenetics is now poised to similarly revolutionize
his contemporary, Italian Nobel laureate Camillo Golgi. knowledge of human development and its perturbations,
The slides revealed the microanatomic structure of the by locating physical points of connection between genes
brain, consisting not of the gelatinous syncytium or the and environments and by revealing a molecular embodi-
fibrous reticulum it had been previously assumed to be, ment of geneenvironment interplay in its propagation of
but rather a breathtakingly complex network of individ- disordered or healthy development. By geneenviron-
ual cells, an immense sea of neurons. In addition to the ment (GE) interplay, we refer to the broad co-action of
brains cellular composition, however, Cajals stained genes and environments on developmental and health
histological sections showed a vast assembly of syn- outcomes (Rutter, 2010). The term epigenetics, in its
apses, clearly visible points of communication between invocation of GE interplay, was first used within a
neurons, numbering in the thousands per individual cell. developmental context in a 1942 paper by biologist
This demonstration of the microanatomic, physical nexus Conrad Waddington to reference a field of study
between the billions of neurons in the human brain exploring the processes by which genotypes are func-
became the provenance of much of what we now tionally linked to adult phenotypes. Waddington argued
recognize as 21st-century neuroscience: the understand- that cell differentiation involved an epigenetic canaliza-

Address for correspondence: W. Thomas Boyce, UCSF Division of Developmental Medicine, Department of Pediatrics, 0503, 550 16th Street, 4th
Floor, San Francisco, CA 94158-0503, USA; e-mail:

2014 John Wiley & Sons Ltd

2 W. Thomas Boyce and Michael S. Kobor

tion of development, analogous to marbles rolling down additional, higher levels of chromatin organization also
a landscape of deepening troughs or canals, seeking the exist, ultimately sculpting the typical structure of chro-
surfaces lowest points, within increasingly irreversible, mosomes and condensing the length of DNA to fit into a
or canalized, trajectories. What began as an intuitive cell nucleus. In broader chromosomal regions, often
metaphor describing crudely observable cell differentia- visible in microscopy stains as chromosomal bands,
tion has become in the years since 1942, and chromatin can be loosely or tightly structured (called
exponentially within the last two decades a science euchromatin and heterochromatin, respectively, and also
occupying the burgeoning epicenter of a new develop- illustrated in Figure 2), thus offering either straightfor-
mental biology. As illustrated in Figure 1, concepts of ward or more difficult physical access of the transcrip-
epigenetic development, as well as an explicit, molecular tional enzyme, RNA polymerase II, to gene promoter
science of epigenesis, have dramatically flourished in the and coding regions. Chromatin configuration is con-
years since Waddingtons observation, with nearly trolled physicochemically by the placement or removal of
20,000 publications bearing the key words epigenetic, chemical tags i.e. epigenetic marks, such as methyl,
epigenesis or epigenetics and listed in PubMed over acetyl, phosphate, or ubiquitin groups on the DNA or
the first four years of the current decade. histone proteins (also shown in Figure 2).

Chromatin structure DNA methylation, histone modification, and microRNA

More recently, epigenetics (from the Greek root epi, The most highly studied and best characterized epige-
meaning upon or over) has been defined as the netic mark, DNA methylation, involves a direct covalent,
structural adaptation of chromosomal regions so as to chemical modification of a cytosine base lying sequen-
register, signal or perpetuate altered [gene] activity states tially adjacent to a guanine base (thus a CpG dinucleo-
(Bird, 2007). Epigenetic mechanisms change gene activ- tide); such methylation is a relatively stable epigenetic
ity or expression by altering chromatin organization, tag, catalyzed by a group of enzymes called DNA
without modifying the genetic code of the DNA (Mea- methyltransferases (DNMTs). CpG methylation can
ney, 2010). Such a definition carries with it the implica- occur during any stage of the cellular life cycle and,
tion that the epigenome is a responsive overlay on the depending upon genomic context, can impede or foster
genome itself, possibly buffering or moderating genetic gene expression or leave it unchanged (Klengel, Pape,
variation through up- and down-regulation of gene Binder & Mehta, 2014). CpG dinucleotides are relatively
transcription. Chromatin is the physical packaging of infrequent in the genome, and areas of comparatively
DNA within chromosomes, in a structural configuration high CpG content have been termed CpG islands
shown in Figure 2. Its most basic unit is the nucleosome, (Illingworth & Bird, 2009). CpG islands tend to be
which comprises 147 base-pairs of DNA wrapped hypomethylated compared to other CpG sites and are
around a histone protein octamer. found associated with approximately 70% of known gene
The 30 million nucleosomes in each cell effectively promoters, i.e. the regulatory, non-coding portion of a
reduce the length of DNA from 2 meters to a packageable gene that plays a role in transcription control (Saxonov,
28 centimeters, thereby appearing in electron microscopy Berg & Brutlag, 2006; Illingworth & Bird 2009).
as a configuration resembling beads on a string Promoter DNA methylation (often in CpG islands)
(each bead constituting a single nucleosome). Several and gene body DNA methylation generally show

Figure 1 Percentage of words epigenetic, epigenesis or epigenetics within Google Ngram English books corpus, 18002000
(; and number of PubMed citations with same three search terms, 19602014.

2014 John Wiley & Sons Ltd

Geneenvironment interplay 3

Histone acetylation

DNA methylation

Figure 2 The chromatin packaging of DNA wound around histone protein octamers, like beads on a string (Darryl Leja, National
Human Genome Research Institute).

opposite associations with gene expression. High levels Ramos, 2014). Demethylation can occur actively or
of promoter DNA methylation are frequently linked with passively (Gibbs, van der Brug, Hernandez, Traynor,
diminished expression, but in the gene body, high Nalls et al., 2010; Bell, Pai, Pickrell, Gaffney, Pique-Regi
methylation is more often coupled with augmented et al., 2011; Lam et al., 2012), and active demethylation
expression (Kass, Landsberger & Wolffe, 1997; Jones, has been shown to occur via hydroxymethylation of CpG
1999). Importantly, recent work in larger cohorts, on dinucleotides (Tahiliani, Koh, Shen, Pastor, Bandukwala
inter-individual variation, has shown that these princi- et al., 2009; Jones, Fejes & Kobor, 2013). The hydrox-
ples hold only when looking across all genes within an ymethylation mark has also been hypothesized to have a
individual. When comparing a single gene across indi- functional role outside of demethylation, but as it is
viduals, the relation between methylation and expression present in significant amounts only in neural and
could be negative, positive, or null (Lam, Emberly, pluripotent cells, this function is not likely to be
Fraser, Neumann, Chen et al., 2012; Gutierrez-Arcelus, important in non-neural tissue (Lister, Pelizzola, Dowen,
Lappalainen, Montgomery, Buil, Ongen et al., 2013). Hawkins, Hon et al., 2009; Globisch, Muenzel, Mueller,
Such associations are therefore not straightforward, and Michalakis, Wagner et al., 2010).
concomitant gene expression is an important additional In a second form of epigenetic mark, post-transla-
component of epigenetic analysis. tional modifications of the nucleosomes histone pro-
DNA methylation, with its accompanying down-reg- teins are reversible chemical tags on the N- and C-
ulation of gene expression, has only recently become terminal histone tails, with acetylation serving a
viewed as reversible and thus a potential mechanism of permissive role in gene transcription through a relax-
developmental plasticity (Wu & Zhang, 2014) and a ation of chromatin structure. Methylation of CpG
useful target of pharmacological interventions (Bojang & sites, which occurs disproportionately in gene promoter

2014 John Wiley & Sons Ltd

4 W. Thomas Boyce and Michael S. Kobor

regions or near transcription start sites in the DNA These three distinctive forms of epigenetic processes
sequence, results either in an interference with tran- each equipped to regulate and/or signal levels of gene
scription factor binding or in the recruitment of expression are not entirely independent, however, and
methyl-DNA binding proteins, which in turn bring recent research has shown how histone methylation can
histone deacetylases (HDACs) to the site. The deacet- suppress nearby DNA methylation and how small
ylation of a histone protein causes, in turn, a ncRNAs and CpG methylation can reciprocally influ-
compaction of chromatin structure and a consequent ence each others presence and effects (Klengel et al.,
physical impediment to gene transcription. DNA 2014).
methylation and histone deacetylation thus work
hand-in-hand to allosterically down-regulate the tran-
Epigenetic differentiation of cells in embryogenesis
scriptional expression of a gene.
Because of the number, complexity and interdepen- These processes of transcriptional control play an
dence of histone modifications (14 distinct marks essential role, through the sequenced complexities of
occurring in 100 sites on histone proteins), the concept embryogenesis, in the ontogeny of a human body
of a histone code or histone language has been comprising ~200 different histological cell types, each
advanced, in which differing combinatorial patterns of containing the same genome, with the same DNA
modifications would drive certain transcriptional and sequence. This proliferation of cellular diversity from a
epigenomic states (Strahl & Allis, 2000; Bridi & Abel, genomic singularity occurs during post-conceptual
2013). Such a pattern-language has been linked, more- development by virtue of a highly regulated, epigenetic
over, to neuronal storage and the expression of memories calibration in the expression of the ~20,00025,000
and behaviors and may be associated, as well, with protein-coding genes. Of the ~28 million CpG sites in
neuropsychiatric conditions such as cognitive impair- the human genome (a smaller subset of the genomes
ments, schizophrenia and depression. Chronic stress ~one billion cytosine bases), as many as 6080% are
down-regulates brain-derived neurotrophic factor methylated within somatic cells (Wu & Zhang, 2014),
(BDNF) expression, for example, through histone acet- and during mitosis, the pattern of DNA methylation is
ylation at the BDNF gene promoter region, an effect replicated in daughter cells, thereby maintaining cellu-
that is reversible with an antidepressant medication lar transcriptional memory. Some deviation from exact
(Shirayama, Chen, Nakagawa, Russell & Duman, 2002). replication occurs for both epigenetic marks and DNA
As a consequence of such observations, pharmaceutical sequence, but the rate of stochastic errors in the
HDACs have become a target of new drug discovery former is estimated to exceed that of the latter by a
efforts in the treatment of cancer, neurological and factor of three (Petronis, 2010). Gene transcription,
psychiatric disorders. and a cells histological fate, is thus controlled by
A third, newly discovered epigenetic mechanism is the epigenetic marks acquired as part of the cell differen-
expression of small, non-coding RNAs (ncRNAs) that tiation process and are dependably reproduced, to a
interfere with the expression of specific genes by asso- remarkable degree, during DNA replication and cell
ciating with DNA to promote compacted heterochro- division.
matin formation or by fostering the degradation of Early embryologic processes also depend critically
transcribed messenger RNA (Mattick, 2010). Such upon the epigenetic programming that underlies cell
ncRNA molecules are highly enriched in the mammalian differentiation and development (Smith & Meissner,
brain. As with other epigenetic modifications, however, 2013). As depicted in Figure 3, methylation of DNA
ncRNAs can serve multiple functions, including the from primordial germ cells undergoes a global reversal,
activation or repression of gene expression, and have followed by an extensive, gamete-specific re-methylation
been linked to several disorders of cognition and process and complex epigenetic remodeling (Strachan &
behavior. For example, in Fragile X syndrome, a herita- Read, 2011). Genomic imprinting, which occurs in
ble disorder causing mental retardation and autistic-like nearly 400 human genes, involves epigenetic modifica-
behavior, a large, trinucleotide repeat sequence in the tions of either DNA or histone proteins acquired from
FMR1 gene coding for Fragile X mental retardation one parental gamete, resulting in only the other parental
protein (FMRP) results in its hypermethylation and gene copy being expressed. After the two parental
transcriptional silencing. As a translation repressor, pronuclei from egg and sperm have fused, the early
ncRNAs are thought to play an additional role in zygotic genome undergoes a second, expansive demethy-
FMRP expression by targeting FMR1 transcripts, lation. Starting at the blastocyst stage of embryogenesis,
impairing synaptic plasticity, and leading to mental there is another genome-wide, de novo methylation
retardation. process that establishes and maintains distinctive cell

2014 John Wiley & Sons Ltd

Geneenvironment interplay 5

Figure 3 DNA methylation/demethylation during embryogenesis (Strachan & Read, 2011).

lineages, giving rise to the broad array of somatic cell Widespread and regionalized shifts in DNA methyla-
types, as well as the trophoblastic cell lines that form the tion and histone modifications have been shown to take
placenta and other reproductive structures. This phasic, place concurrently with key phases of normal brain
genome-wide erasure and reconstitution of methylated development, such as synaptogenesis (Lister, Mukamel,
cytosine nucleotides, at the dawn of embryogenesis, Nery, Urich, Puddifoot et al., 2013; Shulha, Cheung,
creates a kind of genomic tabula rasa upon which the Guo, Akbarian & Weng, 2013), and in association with
epigenetic reprogramming of cellular and organismic specific disorders of development and mental health
diversity can be written. The epigenetic profile of a cell (Champagne, 2010; Toyokawa, Uddin, Koenen & Galea,
then becomes the mechanism whereby a liver cell 2012; Kofink, Boks, Timmers & Kas, 2013). Indeed, it
remembers that it is a liver cell over an entire human appears that the entire process of mammalian neurode-
lifespan. velopment requires a precisely coordinated sequence of
epigenetic events involving genomic methylation and
demethylation in order to produce and spatially locate
Neurodevelopment, dynamic variation and an
functionally distinct populations of neurons and glia
epigenetic paradox
cells (Wu & Zhang, 2014). Further, Qureshi and Mehler
Interestingly, however, the very same processes that (Qureshi & Mehler, 2010) and Hodes (Hodes, 2013) have
instantiate this ontogenetic stability in cell differentiation reviewed evidence that DNA methylation, histone mod-
are those that also ensure dynamic variation in transcrip- ifications and ncRNAs may all be implicated in the
tional activity in response to environmental signals and known sexual dimorphisms arising during brain devel-
conditions. Cellular activity, such as the depolarization opment, which may underlie differential susceptibility
of neurons during neural circuitry activation, induces among males and females to various forms of psycho-
moment-to-moment shifts in gene activation or deacti- pathology. Epigenetic links to neurodevelopmental
vation. In one example elucidated by Lubin et al. abnormalities and disorders of mental health include
(Lubin, Roth & Sweatt, 2008), exposure to a fear the distinctive methylation patterns found within hun-
learning procedure in rats produced an activity-depen- dreds of gene loci, among patients with autism spectrum
dent, dynamic regulation of the BDNF gene in the disorder and other neurodevelopmental syndromes
hippocampus by DNA methylation, in the absence of (Shulha, Cheung, Whittle, Wang, Virgil et al., 2012;
cell division, and in response to environmental influ- Berko, Suzuki, Beren, Lemetre, Alaimo et al., 2014).
ences. Such epigenetic regulation is ubiquitous in the Children with Down Syndrome, with their triplication of
brain and is responsible for a variety of complex neural chromosome 21, should theoretically have 50% more
functions, such as memory formation, learning and the expression of those chromosomal genes, but their actual
calibration of stress response circuitry (Bohacek, Gapp, transcription varies from that expectation, suggesting
Saab & Mansuy, 2013). that DNA methylation, histone modification and

2014 John Wiley & Sons Ltd

6 W. Thomas Boyce and Michael S. Kobor

possibly other epigenetic events may be involved (Dek- during embryogenesis (Lowe et al., 2013). Since different
ker, De Deyn & Rots, 2014). Such differences in tissues show distinctive epigenetic patterns, methylome
epigenetic adaptation might give rise to the cognitive comparisons allow an assessment of how representative
deficits that co-occur in Down Syndrome, rendering central tissues may be of their peripheral counterparts.
them potentially amenable to pharmacological interven- One area of particular interest is epigenetic variability,
tions targeting the epigenome. i.e. determining whether the amount of variation
Thus, in an epigenetic paradox, the molecular between individuals found in peripheral tissues matches
sources of longitudinal cellular stability, by which the variation within central tissues.
histological differentiation occurs in early embryogene- Second, when comparing a specific tissue methylation
sis, are co-opted to subserve the organisms finely tuned, profile across individuals, differences in the cellular
short- and long-term responsivity to shifting environ- composition of the tissue sample can substantially affect
mental circumstances. Such responsivity is neither fully profile differences between the individuals (Jaffe &
environmental nor fully genetic in origin, but rather a Irizarry, 2014). This can be corrected by directly mea-
synthesis, a convergence of gene and context. For a more suring sample tissue composition, by using DNA meth-
complete treatment of the molecular and developmental ylation profiles themselves to back-predict underlying
processes summarized here, the reader is referred to cellular composition, or by using methods that correct
books by Allis and colleagues (Allis, Jenuwein, Reinberg for composition without the actual measurements (see
& Caparros, 2007), Gilbert and Epel (Gilbert & Epel, e.g. Houseman, Accomando, Koestler, Christensen,
2009), and Sweatt and colleagues (Sweatt, Meaney, Marsit et al., 2012; Lam et al., 2012; Zou, Lippert,
Nestler & Akbarian, 2013). Heckerman, Aryee & Listgarten, 2014). Based on
existing literature, it is likely that the within-individual
differences in DNA methylation between tissues trumps
Tissue specificity of epigenomic variation
inter-individual differences in the same tissue, rendering
Closely related to this paradoxical colocation of cellular between-tissue differences the main drivers of DNA
stability and change within a single epigenomic structure methylation variability (Davies, Volta, Pidsley, Lunnon,
is the dilemma of how to study dynamic epigenetic Dixit et al., 2012; Lam et al., 2012; Ziller, Gu, Muller,
change against a backdrop of systematic, tissue-specific Donaghey, Tsai et al., 2013).
differences in epigenetic profiles. How can the epigenetic Although the interpretive challenges of epigenetic
modifications associated with early adversity in hippo- studies in peripheral tissues are formidable, recent,
campal neurons arguably a logical target tissue in non-human primate research by Provencal et al. (Pro-
which to search for such effects be reliably discerned vencal, Suderman, Guillemin, Massart, Ruggiero et al.,
from those found in buccal epithelial cells (BECs)? What 2012) compared DNA methylation levels and sites
can chromatin modifications found in peripheral blood between cells from the prefrontal cortex (PFC) and T
mononuclear cells (PBMCs) possibly reveal of the lymphocytes in peripheral blood and found both tissue-
epigenetic marks in brain circuitry controlling the HPA specific differences and other methylation patterns that
or autonomic nervous system (ANS) circuitry? co-varied with early rearing conditions (mother- versus
In the context of human epigenetic studies, histolog- peer-reared) across cell types. Some rearing condition
ical differences in DNA methylation pose a difficulty in effects were also found only in T cells, consistent with
two ways. First, since in many cases the tissue of interest prior observations that early adversity has an immune
for a particular condition may not be available (e.g. brain component unlikely to co-occur in brain. In sum,
tissue in PTSD), post-mortem, surgical or surrogate important caveats about the limitations of peripheral
tissues are often employed. Post-mortem tissue has been epigenomic measures notwithstanding, there is likely
useful in some studies (e.g. McGowan, Sasaki, DAles- much to be learned from the epigenetic variation
sio, Dymov, Labonte et al., 2009), but in general, for observable within peripheral tissues, perhaps even
large-scale studies, only surrogate tissue is available among human children.
(Horvath, Zhang, Langfelder, Kahn, Boks et al., 2012;
Lowe, Gemma, Beyan, Hawa, Bazeos et al., 2013). In
these cases, researchers examine peripheral tissues, such Geneenvironment interplay and epigenetic
as BECs or PBMCs, for associations with phenotypes processes
manifesting in central tissues. Some have argued, for
example, that the DNA methylome of BECs may be GE interplay has emerged as a promising point of
more reflective of epigenomic states in brain structures, origin in studies of divergent developmental trajectories
as a consequence of their common ectodermal origin and the emergence of mental disorder. Such interplay is

2014 John Wiley & Sons Ltd

Geneenvironment interplay 7

thought to comprise at least three categories of pro- within the samples of modest size, where such effects
cesses: (a) geneenvironment correlation (rGE), (b) G9E have largely been found (e.g. Duncan & Keller, 2011),
interaction (G9E), and (c) epigenetic, chromatin mod- attributing most G9E reports to a combination of Type
ifications. First, rGE refers to a genetic biasing of I error and publication bias. Other critiques of these
environmental exposures, in the sense that individuals dissenting, meta-analytic reviews, however, have pointed
can be predisposed, based on genetic background, to to selectivity in the choice of studies, unwarranted
select, alter and/or generate specific categories of expe- statistical assumptions, and a failure to consider the
riences e.g. the behaviorally inhibited childs choice of expectation of G9E interaction on biological grounds
less challenging or intensive social environments. Second, (Rutter, Thapar & Pickles, 2009). A 2010 review of 40+
G9E designates genetic or environmental effects on studies of G9E interactions involving the serotonin
outcomes that are conditional upon each other e.g. the transporter polymorphism revealed not only predomi-
effects of genetic variants becoming apparent only in the nantly strong (though not unmixed) confirmation of the
presence of specific environmental conditions, or envi- effect, but the appearance, as well, of a large number of
ronmental influences being revealed only among indi- convergent neuroscience experiments supporting the role
viduals of a particular genotype. Third, as discussed of the short risk allele in generating neural sensitivity to
above, epigenetic marks controlling the structural con- negative and stressful environments (Caspi, Hariri
figuration of chromatin are the means by which envi- Holmes, Uher & Moffitt, 2010). Further, in a 2011
ronmental signals guide and adjust gene transcription to meta-analysis by Karg et al. (Karg, Burmeister, Shedden
maximize adaptation, fitness and health. & Sen, 2011) a more inclusive strategy merging study
The exploration of these three domains of GE findings at the level of significance testing, rather than
interplay has become a prolific and engaging area of raw data, allowed the consideration of 54 studies; the
biomedical and social science research. Indeed, GE authors concluded that there is strong evidence for the
interplay research holds at least implicit promise for moderation of the stressdepression association by the
illuminating one of the oldest and deepest mysteries of 5HTTLPR polymorphism.
human experience, i.e. how individual susceptibility and Despite legitimate concerns for the replicability of
social conditions work together at the behavioral, G9E interaction reports, some such effects have been
physiologic, neural, cellular and molecular levels to replicated in independent samples (van Winkel, Peeters,
initiate and sustain disorders of development, behavior van Winkel, Kenis, Collip et al., 2014), and new evidence
and health. In animal models from fruit flies (Burns, for G9E interactions continues to accrue. Zohsel et al.
Svetec, Rowe, Mery, Dolan et al., 2012) to rats (Meaney, (Zohsel, Buchmann, Blomeyer, Hohm, Schmidt et al.,
2001) and non-human primates (Barr, Newman, Lindell, 2014) reported an interaction between the 7-repeat allele
Shannon, Champoux et al., 2004a), new evidence has of the DRD4, dopamine receptor gene and maternal
accumulated that GE interplay plays a role in the reports of prenatal stressors in predicting risk for
genesis of species-typical and deviant behavior. Though diagnoses of conduct disorder or oppositional defiant
such evidence had been long anticipated, it was only a disorder in early adolescence. Drury et al. (Drury,
decade ago that reports began to emerge documenting Theall, Smyke, Keats, Egger et al., 2010), from the
G9E interactions in the longitudinal prediction of Bucharest Early Intervention Project, identified a G9E
human developmental and health outcomes. The papers interaction in which children who remained institution-
of Caspi, Moffitt and colleagues (Caspi, McClay, Mof- alized and were carriers of the met allele of the COMT,
fitt, Mill, Martin et al., 2002; Caspi, Sugden, Moffitt, catechol-O-methyltransferase gene had significantly
Taylor, Craig et al., 2003) from the Dunedin Multidis- higher mean levels of depressive symptoms. COMT
ciplinary Health and Development Study revealed sta- degrades catecholamine neurotransmitters, such as
tistical interactions between early environmental dopamine and norepinephrine, and has been implicated
conditions (e.g. child maltreatment and stressful life in risk for major depression. Broekman and colleagues
events) and functional gene polymorphisms (e.g. the (Broekman, Chan, Goh, Fung, Gluckman et al., 2011),
MAOA, monoamine oxidase A, and 5HTT, serotonin examining socioemotional development in children from
transporter, genes) in the prediction of antisocial behav- a cohort study in Singapore, found that specific allelic
ior and depression/suicidality. variants in three genes involved in serotonergic func-
Questions were raised regarding the 5HTTLPR 9 tioning (TPH2, SCL6A4, and HRT2A) moderated the
stressful events interaction in a meta-analysis of reports influence of birth weight on internalizing symptoms at
that followed the Caspi findings (Risch, Herrell, Lehner, 812 years of age. SNPs in each of the three serotonin-
Liang, Eaves et al., 2009), and some have more recently related genes were associated with a significant reduction
dismissed efforts to identify significant G9E effects in symptoms, but only among children occupying the

2014 John Wiley & Sons Ltd

8 W. Thomas Boyce and Michael S. Kobor

third quartile of birth weight (i.e. the quartile immedi- new knowledge of how multiple genes with additive or
ately above the median). Further, in another form of multiplicative effects, each with incremental influences,
G9E interaction detectable in twin samples, Tucker- are assembled into functional genetic networks that
Drob et al. (Tucker-Drob, Rhemtulla Harden, Turkhei- interact with social environmental conditions to produce
mer & Fask, 2011) used data from the Early Childhood important phenotypic disorders (Gratten, Wray, Keller
Longitudinal Study to show that genetic variation in & Visscher, 2014).
cognitive ability depends upon reciprocal, developmen-
tally moderated interactions between child and environ-
Epigenetics and the mediation of G9E interactions
ment, such that by age 2 years, genetic effects on Bayley
Short Form scores were larger for children being raised Rothman and Greenland (Rothman & Greenland, 2005)
in higher socioeconomic status (SES) homes. have argued that, in effect, all disorders of health and
Beyond these prospective, observational studies, true development are uniformly both genetic and environ-
experimental evidence for G9E interaction has emerged mental, in the sense that virtually all endpoints depend
from random allocation trials with non-human primates. upon mutually interactive influences of both. Among the
Suomi and colleagues, for example, have demonstrated, most intriguing recent discoveries in support of such a
in an experimental model with rhesus macaques how perspective implicate epigenetic processes as possible
early rearing conditions, within either mother- or peer- molecular mechanisms for G9E at the population level.
reared groups, interact with the 5HTTLPR, serotonin It is known, for example, that maltreatment of children
transporter promoter polymorphism to predict ACTH bears important but complex linkages to dysregulation
expression during separation stress (Barr, Newman, of the hypothalamic-pituitary-adrenocortical (HPA) axis
Shannon, Parker, Dvoskin et al., 2004b). Such interac- (Tarullo & Gunnar, 2006), to pro-inflammatory shifts in
tions appear to apply even in regard to the timing of cytokine signaling pathways (Danese, Pariante, Caspi,
normative developmental events, as shown in another Taylor & Poulton, 2007), and to long-term changes in
rhesus model in which social dominance status during epigenetic and gene expression signatures within key,
development interacted with the 5HTTLPR polymor- stress-responsive neural structures (McGowan et al.,
phism to predict the timing of sexual maturation (Wilson 2009). Exploring the known interaction between child-
& Kinkead, 2008). Specifically, subordinate female hood maltreatment and an allelic variant in the FKBP5,
monkeys carrying at least one copy of the short FK506 binding protein 5 gene in predicting adult PTSD,
promoter variant had significant delays in pubertal for example, the Binder laboratory found a mediating
sexual maturation. molecular event involving demethylation of a CpG site
Thus far, such experimental G9E effects have not within an intron of the risk allele. FKBP5 codes for a so-
been extended to human samples, though there are no called chaperone protein that alters the function of the
prohibitive reasons why causal evidence could not also glucocorticoid receptor (GR) and impedes the translo-
be derived from human groups, using study designs with cation of the GR-glucocorticoid complex into the cell
random assignment by genotype. As noted by Kraemer nucleus. The demethylation event, which can occur only
(Kraemer, 2012), the research design and mathematical during an early critical period and results in a persistent
modeling difficulties inherent in detecting and interpret- activation of FKBP5, then primes the risk allele carrier
ing the cooperation of genes and environments create a for stress-induced over-production of the chaperone
perfect storm of methodological challenges. That protein, suppression of GR function, HPA axis dysre-
tempest occupies the dead-center, however, of the pos- gulation and a consequent augmentation of risk for
sibly most fertile and promising arena of contemporary PTSD. Though Binders group provides compelling
biomedical research: i.e. how genes and environments laboratory evidence for a molecular, epigenetic mecha-
work together to undermine health. The search for nism mediating the FKBP5 9 maltreatment interaction,
reliable G9E interactions may be abetted by the devel- it would be important in future epidemiologic work to
opment of both empirical evidence that polygenic risk estimate the portion of variance in methylation states
scores associated with developmental phenotypes can that is attributable to the interaction. Epigenetic marks
discern promising new SNP targets (Rietveld, Medland, (whether DNA methylation/demethylation or post-trans-
Derringer, Yang, Esko et al., 2013) and computational lational histone protein modifications) are likely only one
models suggesting that genome-wide association study of several molecular mechanisms controlling gene
(GWAS) approaches to G9E discovery may be more expression, but this empirically driven intersection of
promising than candidate SNP by SNP searches (Mari- the fields examining genetic and epigenetic variation
gorta & Gibson, 2014). Indeed, in fields such as offers one groundbreaking account of how and under
psychiatric genomics, the way forward appears to lie in what conditions G9E interactions arise.

2014 John Wiley & Sons Ltd

Geneenvironment interplay 9

While Klengel et al. shed new light on the epigenetic Taken together, these recent observations suggest that
mechanism of a G9E interaction, findings from the much of developmental variation, and in particular that
Holbrook laboratory by Teh et al. (Teh, Pan, Chen, involving risks for or manifestations of developmental
Ong, Dogra et al., 2014) make essentially the same psychopathology, may be linked to interactions between
inference, but from the reverse direction: showing how polygenic and environmental differences. Further, emerg-
neonatal human methylomes are affected by both ing evidence implicates epigenetic processes as molecu-
individual, intrauterine exposures and fixed genetic lar-level mechanisms by which such interactions may
variation, i.e. G9E interactions. Specifically, surveying occur. Thus, the statistical interactions identified
the methylomes of over 200 Singaporean newborns, the between allelic variants and risk-engendering early social
investigators found over 1400 punctate genomic regions environments may be rooted in and attributable to
that were highly variable in methylation status across differences in DNA methylation, post-translational his-
individuals locations termed variably methylated tone modifications, or ncRNAs resulting in inter-
regions (VMRs). In assessing the environmental (e.g. individual variation in the transcription of genes linked
birthweight as a proxy for nutrition, smoking, maternal to pathological phenotypes or endophenotypes.
depression, and maternal body mass index), genetic and
G9E origins of VMRs, G9E interactions were found to
account for 75% of VMRs, through a combination of An epigenetic embedding of early deprivation,
intrauterine environmental signatures on the fetal epi- adversity and developmental risk
genome and the physicochemical effects of sequence
variation on a CpG sites propensity for methylation. In
Adversity and epigenetic marks
fact, no VMRs were best accounted for by environmen-
tal linkages alone, independent of gene sequence vari- Among such endophenotypes are the patterns of epige-
ation. These findings reveal an increasingly convergent netic modifications that attend early exposures to depri-
picture of how genetic variation in DNA sequences, vation, maltreatment and adversity. Recent evidence, from
naturally occurring (and sometimes pernicious) varia- both experimental animal models and human observa-
tion in social environmental exposures, and their tional studies, reveals reliable deprivation- and stress-
molecular-level interactions operate, through chromatin related differences in DNA methylation and histone
modifications and other transcription regulatory pro- modification that are developmentally timed and affect
cesses, to produce combinatorial disturbances in devel- stress-responsive, neuroendocrine pathways with known
opmental and health endpoints. linkages to developmental psychopathology (Monk, Spic-
Lastly, Kobor and colleagues, in a series of studies er & Champagne, 2012). In a set of transformative studies,
examining epigenetic and allelic variation within human for example, Meaney and colleagues (Weaver, Diorio,
populations, have shown that DNA methylation profiles Seckl, Szyf & Meaney, 2004, Meaney 2010) used naturally
are highly divergent between human populations and occurring differences in maternal licking and grooming of
that the sources of such divergences involve differences in rat pups to demonstrate how low maternal caretaking up-
allelic frequencies and complex epistatic1 and G9E regulates pups long-term HPA reactivity through
interactions (Fraser, Lam, Neumann & Kobor, 2012). decreases in hippocampal GR (NR3C1) expression and
Examining over 14,000 genes in 180 different cell lines serotonergic tone. Maternal licking and grooming triggers
from European and African samples, they found popu- increases in serotonin (5-HT) expression in the hippo-
lation-level differences in DNA methylation near tran- campus, and activation of the 5-HT receptor induces a
scription start sites in over a third of genes. Further transcription factor, nerve growth factor-inducible pro-
analyses indicated that these methylation differences tein-A (NGFI-A). Maternal care also facilitates NGFI-
were mostly attributable to differences in allele frequen- As association with the exon 17 GR promoter by
cies. Several other groups have similarly documented demethylating a CpG site located in the promoters
such epigenetic dissimilarities between human popula- NGFI-A binding region. Low licking and grooming is
tions (Adkins, Krushkal, Tylavsky & Thomas, 2011; thus linked to increased DNA methylation and decreased
Heyn, Moran, Hernando-Herraez, Sayols, Gomez et al., histone acetylation within the NR3C1 gene, resulting in
2013; Moen, Zhang, Mu, Delaney, Wing et al., 2013). diminished expression of GR, an up-regulation of corti-
cotropin releasing hormone (CRH) secretion, and greater
activation of the HPA axis.
Interactions in which the phenotypic effect of a given gene is In other rodent work, early infant maltreatment has
moderated by the effects of another gene or network of genes (i.e. G9G been linked to an enduring, increased methylation of the
interactions). BDNF, brain-derived neurotrophic factor gene, reducing

2014 John Wiley & Sons Ltd

10 W. Thomas Boyce and Michael S. Kobor

BDNF expression (Roth, Lubin, Funk & Sweatt, 2009). developmentally and immunologically active genes
Chronic, variable stress during the first trimester of (Tobi, Lumey, Talens, Kremer, Putter et al., 2009). Other
pregnancy in mice results in heightened corticosterone research with institutionalized children 710 years of age
expression and increased depressive behavior in the reported whole genome hypermethylation, compared to
offspring, linkages that are mediated by sex-specific parent-reared controls (Naumova, Lee, Koposov, Szyf,
differences in DNA methylation in the CRH and NR3C1 Dozier et al., 2012). Oberlander and colleagues (Ober-
genes (Mueller & Bale, 2008). Maternal behavior, neglect lander, Weinberg, Papsdorf, Grunau, Misri et al., 2008)
and abuse by no means affect only the epigenetic states reported increased NR3C1, GR gene methylation among
of genes involved in HPA regulation, however, as over infants born to mothers with high depressive symptoms
900 genes are differentially expressed in the hippocampus during the third trimester of pregnancy, and Radtke
as a result of low versus high maternal care (Weaver, et al. (Radtke, Ruf, Gunter, Dohrmann, Schauer et al.,
Meaney & Szyf, 2006). There is also experimental 2011) derived similar findings from the leukocytes of
evidence, summarized by Korosi et al. (Korosi, Naninck, adolescents whose mothers were exposed to intimate
Oomen, Schouten, Krugers et al., 2012) that the effects partner violence during pregnancy. Romens et al. (Ro-
of early life stress on adult neurogenesis in the rat may be mens, Svaren & Pollak, 2014), in a community sample of
mediated by various components of the epigenetic early adolescents, found similarly increased DNA meth-
molecular machinery. ylation of the GR promoter among children who had
Evidence for early social adversity effects on epigenetic been physically abused, compared to peer controls, and
states has now extended to non-human primates, as well. similar results have been reported in hippocampal cells
Suomi and others (Provencal et al., 2012; Tung, Barreiro, from suicide victims with a history of child abuse
Johnson, Hansen, Michopoulos et al., 2012) have shown (McGowan et al., 2009; Sasaki, de Vega & McGowan,
that social dominance rankings and rearing conditions are 2013). Ouellet-Morin and colleagues (Ouellet-Morin,
associated with epigenetic and gene regulatory variation Wong, Danese, Pariante, Papadopoulos et al., 2013)
in the immune system of rhesus macaques. Mother- versus found more extensive DNA methylation of the serotonin
peer-rearing results in differential methylation patterns in transporter (SERT) gene among bullied monozygotic
the infants prefrontal cortical neurons and T lympho- twins than in their non-bullied co-twins, and, in a sample
cytes. Further, such patterns in response to rearing of healthy adults, other investigators found reports of
conditions were not randomly distributed across the parental loss, maltreatment, and impaired parental care
genome but showed a structural organization targeting to be associated with GR methylation status (Tyrka,
specific cellular functions. Cole and colleagues (Cole, Price Marsit, Walters & Carpenter, 2012).
Conti, Arevalo, Ruggiero, Heckman et al., 2012), also Despite these multiple reports of heightened, adver-
working with infant rhesus macaques, found up-regulated sity-related methylation in the human and animal GR
expression of genes involved in pro-inflammatory cyto- gene promoter, findings in this regard are not uniformly
kine signaling and T-cell activation among peer-reared positive in either animals or humans. Methylation
monkeys, as well as suppressed expression of other genes differences are often quite modest in magnitude, and
with roles in antimicrobial defenses. Kinnally and col- confounds, such as the proportions of individual cell
leagues (Kinnally, Feinberg, Kim, Ferguson, Leibel et al., types in peripheral blood, have frequently not been taken
2011), in a study of bonnet macaques randomly assigned into account. Witzmann et al. (Witzmann, Turner, Mer-
to an early, stressful variable foraging demand condition, iaux, Meijer & Muller, 2012), for example, in an early
found associations among developmental exposures to chronic stress model in rats, confirmed some DNA
feeding adversity, behavioral stress reactivity, enhanced demethylation in the GR promoter, but in contrast to
whole genome methylation, and epigenetic modification Weaver (Weaver et al., 2004), was unable to detect
of the 5HTT, serotonin transporter gene. hypomethylation in the NGFI-A recognition site of the
Human studies of early adversity and neglect show GR 17 promoter. In another example, Alt et al. (Alt,
parallel evidence for effects on endophenotypic epige- Turner, Klok, Meijer, Lakke et al., 2010) found that GR
netic marks, in relation to both pre- and post-natal promoter methylation was unchanged in the post-mor-
events. Periconceptual exposures to famine and adversity tem brains of patients with major depression but no early
during the 194445 Dutch Hunger Winter, for example, trauma, whereas McGowan and colleagues (McGowan
were associated with decreased methylation in the IGF2, et al., 2009) had reported GR promoter demethylation in
insulin-like growth factor II gene, which has an the brains of individuals with histories of child abuse. It
important role in developmental biology and growth is thus possible that GR demethylation, as one example
(Heijmans, Tobi, Stein, Putter, Blauw et al., 2008), as of an adversity-related epigenetic modification, is highly
well as differential methylation in a variety of other specific with respect to both the character of the early

2014 John Wiley & Sons Ltd

Geneenvironment interplay 11

adversity encountered, as well as the exposure timing exposures are biologically embedded into epigenetic
(Klengel et al., 2014). processes potentially affecting health risks, and that
Nonetheless, other epidemiologic research at a more broad individual differences exist, also epigenetically
genome-wide level by Borghol et al. (Borghol, Suder- mediated, in the health and developmental consequences
man, McArdle, Racine, Hallett et al., 2012), Essex et al. of stress.
(Essex, Boyce, Hertzman, Lam, Armstrong et al., 2013) Stress-related psychiatric conditions, such as suicidal
and Kobor and colleagues (Lam et al., 2012; Powell, ideation and attempts, have been associated, for example,
Sloan, Bailey, Arevalo, Miller et al., 2013) detected with genetic variation in the molecular machinery of
longitudinal associations between childhood disadvan- epigenetic processes, such as polymorphisms found in the
tage and genome-wide promoter methylation in mid-life, DNMT gene (Murphy, Mullins, Ryan, Foster, Kelly
between parental stress in infancy and differential et al., 2013). In another study of depressed adolescents
methylation in adolescence, and between early socioeco- and their unaffected peers, symptoms were more com-
nomic status and inflammatory gene expression in mon among youth who showed increased buccal epithe-
leukocyte transcriptomes. Such findings, of broad epige- lial cell methylation in the 5HTT, serotonin transporter
nomic differences in DNA methylation status, are gene, along with the short-allele promoter in the same
commensurate with the known cellular effects of stress gene (Olsson, Foley Parkinson-Bates, G. Byrnes, M.
hormones, such as glucocorticoids, which are known to McKenzie et al., 2010). Mehta, Binder et al. (Mehta,
moderate expression of approximately 10% of the Klengel, Conneely, Smith, Altmann et al., 2013) found
genome (Matthews & Phillips, 2012). They are sup- distinctive patterns of whole genome expression and
ported, as well, by reports of real-time, dynamic changes methylation among PTSD patients, with and without
in gene methylation among individuals exposed to histories of childhood abuse. PTSD patients with
laboratory-based, stressful challenges (Unternaehrer, accompanying childhood trauma exposure had 12-fold
Luers, Mill, Dempster, Meyer et al., 2012). Taken increases in DNA methylation, relative to those without
together, the observations offer credible evidence, in such histories. There is increasing evidence for epigenetic
both animal and human systems, for a possibly extensive mechanisms involved in schizophrenia, much of it
interplay among genes, epigenomes, and social environ- focused upon hypermethylation within a large CpG
ments in the genesis of epigenomic endophenotypes. island in the promoter region of the gene encoding reelin,
a glycoprotein found in GABAergic neurons (Tsankova,
Renthal, Kumar & Nestler, 2007). Finally, psychoactive
Psychopathology and epigenetic marks
drugs, such as cocaine and some antipsychotic agents,
The endophenotypes related to early adversity are of have been noted to induce acetylation and other mod-
great salience to human population health because of the ifications of histone proteins (Tsankova et al., 2007). In
abundant social environmental challenges with which sum, studies documenting a variety of pathophysiolog-
millions of young children contend around the globe, ical changes in brain and neural circuitry including
including poverty, war, subordination and bullying, changes in regional structure and function, differences in
maltreatment, parental mental illness or addiction, circuitry activation, molecular dysregulation at the
family dissolution, and exposures to violence in both synaptic cleft, and alterations in intracellular kinetics
home and community. There is now a substantial body and signaling pathways may all have in common
of evidence and a strong, corresponding scientific dysregulatory changes in epigenetic processes that
consensus that such adversity and stress in early life are underlie these fundamental neurobiological features.
also associated with disturbances of childhood mental
health, more disordered developmental trajectories,
Individual variation in epigenetic susceptibility
poorer educational achievements, and lifelong risks of
chronic disorders of health and well-being (see e.g. At the levels of both behavior and biology, however,
Shonkoff, Boyce & McEwen, 2009; Hertzman & Boyce there are dramatic differences in the consequences of
2010; Boyce, Sokolowski & Robinson, 2012). Further, exposures to early adversity, with many children showing
there is related evidence that such experiences of toxic immediate and long-term deficits in health and develop-
stress2 are socioeconomically partitioned, that the ment, while others thrive and survive with apparent
indifference to the challenges they face (Rutter, 2012;
That is, stress involving strong, frequent and/or prolonged exposures
Masten, 2014). The source of this individual variation in
to adversity, without adequate adult support, and sufficient to activate the consequences of toxic stress has been the focus of
neurobiological stress response systems (National Scientific Council on increasing study, since understanding such differences
the Developing Child, 2005/2014). could explain stress-related disorders, shed light on the

2014 John Wiley & Sons Ltd

12 W. Thomas Boyce and Michael S. Kobor

sources of personal resilience and vulnerability, explain colleagues (Bogdan, Agrawal, Gaffrey, Tillman & Luby,
the uneven distribution of disease within human popu- 2014) also found a 5HTTLPR 9 stressful life events
lations, and produce insights leading to more effective interaction in predicting preschool-onset depressive
intervention strategies. The stress diathesis perspective symptoms among 35-year-old children. Again, those
on such differences in stress response held that individ- with the risk allele had either the most or least depressive
uals varied in the impact of adversity by virtue of either symptoms, depending upon level of stressor exposure.
heritable or acquired vulnerabilities to stress and chal- Brett et al. (Brett, Sheridan, Jones, Esteves, Fox et al.,
lenge. More recently, it has become apparent that 2014), examining data from the Bucharest Early Inter-
another, possibly more prevalent form of variability in vention Project, found that the high susceptibility allele
contextual effects is a differential susceptibility to envi- of the ERBB3, neuroregulin gene predicted the largest
ronmental influence, in which a subset of individuals corpus callosum volumes in children randomized to the
appears more sensitive or permeable to the influences foster care condition, but the smallest volumes among
of both negative and positive environmental factors. In a those remaining in orphanages. In a meta-analysis of
now substantial body of literature (e.g. Boyce, Chesney, how negative and positive rearing environments are
Alkon-Leonard, Tschann, Adams et al., 1995; Belsky, linked to developmental outcomes, Bakermans-Kranen-
2005; Boyce & Ellis, 2005; Ellis, Boyce, Belsky, Baker- burg and van IJzendoorn (Bakermans-Kranenburg &
mans-Kranenburg & van Ijzendoorn, 2011), differen- van IJzendoorn, 2011) found that children with the less
tially susceptible children (sometimes referred to as efficient, 7-repeat DRD4, dopamine receptor allele fared
orchid children, in contrast to their more resilient less well in negative environments than their counter-
counterparts, dandelion children) show either the most parts without the genetic risk factor, but also gained
maladaptive or most positive outcomes, depending upon most from positive rearing conditions. These findings
the character of their social environments. and others indicate that allelic variation in DNA
A variety of studies have identified genetic polymor- sequences can itself engender a heightened sensitivity
phisms that appear to function as sources of differential to both negative and positive early social settings.
susceptibility. Bush et al. (Bush, Guendelman, Adler & Within differential susceptibility theory, however,
Boyce, 2014, submitted), for example, have recently greater sensitivities to the character of the social world
shown that the BDNF Val66Met polymorphism confers have also been hypothesized to emerge developmentally
an increased neuroendocrine sensitivity to socioeco- and responsively, via conditional adaptations to the
nomic context, with Met-carriers having the highest social signals of early life (Boyce & Ellis, 2005; Ellis,
and lowest cortisol expression levels, depending upon Essex & Boyce, 2005). Conditional adaptations such as
SES. Consistent with the differential susceptibility the polyphenism in wing pattern and coloration found
hypothesis, Belsky and Beaver (Belsky & Beaver, 2011) among butterflies emerging from chrysalises during
found that an index of risk alleles in five candidate differing seasonal conditions are fitness-augmenting
plasticity genes (DAT1, DRD2, DRD4, 5HTT, and changes in development in response to early environ-
MAOA) moderated the link between parenting quality mental cues (Gilbert & Epel, 2009; Ellis & Bjorklund,
and male adolescent self-regulation. In a sample of 2012; Nederhof & Schmidt, 2012). An important ques-
several hundred African American young adults, Simons tion is therefore whether epigenetic modifications,
et al. (Simons, Lei, Beach, Brody, Philibert et al., 2011) acquired as a consequence of early environmental
found longitudinal evidence for the differential suscep- signaling, might also be linked to differentially suscep-
tibility of aggressive behavior to environmental adversity tible phenotypes. Suderman et al. (Suderman, McGo-
among individuals with the combination of the S-allele wan, Sasaki, Huang, Hallett et al., 2012), for example,
of the 5HTT, serotonin transporter gene and the L-allele studied large, differentially methylated regions centered
of the DRD4, dopamine receptor gene. Similarly, upon the NR3C1, GR gene in the hippocampi of both
Babineau et al. (Babineau, Gordon Green, Jolicoeur- rats and humans experiencing substantially different
Martineau, Minde, Sassi et al., 2014) demonstrated levels or forms of early parental care. The methylation
differential susceptibility among children with the S- profiles of both species were extensively different in
and LG alleles of the 5HTT gene (the latter polymor- individuals receiving high- vs. low-level (rats) or abusive
phism a variant of the L allele with a functional effect on vs. non-abusive (humans) early parental care, with many
mRNA expression similar to that of the S-allele): between-species commonalities in the specific, differen-
children with the risk allele having greater behavioral tially methylated sites. Beach et al. (Beach, Brody, Lei,
and cognitive dysregulation when exposed to prenatal Kim, Cui et al., 2014), in a sample of African American
maternal depression, but greater regulatory capacities in youth from working poor communities, found that
the absence of such prenatal exposure. Bogdan and cumulative socioeconomic adversity and the S-allele of

2014 John Wiley & Sons Ltd

Geneenvironment interplay 13

the 5HTT, serotonin transporter gene interactively pre- plasticity during defined windows of early life (Takesian
dicted promoter region methylation within a group of & Hensch, 2013). A classic example is the deprivation
200+ depression-related genes. Youth with the S-allele amblyopia that occurs in children lacking patterned
had either the highest or lowest levels of depression gene visual stimulation, due to strabismus, cataracts or other
methylation, depending upon levels of exposure to occlusions of vision, during the early development of the
poverty-related stress. Strunk and colleagues (Strunk, brains visual circuitry (i.e. birth to 78 years of age). As
Jamieson & Burgner, 2013) have argued that the Kuzawa and Thayer noted (Kuzawa & Thayer, 2011),
increased susceptibility of infants to infectious agents human adaptation to environmental conditions takes
of disease may be due to differential methylation of place at a variety of timescales, ranging from homeostatic
immune-regulating and other developmentally salient changes that can occur over seconds or minutes to
genes. Finally, the Binder laboratory (Klengel, Mehta, developmental plasticity present over months or years, to
Anacker, Rex-Haffner, Pruessner et al., 2013) has dem- conserved genetic changes that operate on a timescale of
onstrated a differential susceptibility of individuals millennia. Wright and Christiani (Wright & Christiani,
bearing the AG/AA risk allele of the FKBP5 gene. 2010) point out that the critical periodicity of growth and
Such individuals have either higher or lower rates of development occurs as a consequence of the timing and
adult PTSD, conditional upon childhood exposures to sequencing of important neurodevelopmental processes,
sexual and/or physical abuse. Further, the molecular such as cell migration, synaptic proliferation and prun-
process by which this epidemiologically observed inter- ing, changes in receptor density and axonal myelination.
action occurs is mediated through DNA demethylation There is evidence, for example, that the developing brain
in the glucocorticoid response elements of FKBP5. These is especially vulnerable to the deleterious effects of
observations are among the first to show how chromatin chemical exposures during early developmental periods,
modification and epigenetic marks may constitute actual rendering children more liable than adults to toxic injury
molecular mechanisms for a differential susceptibility to during critical periods of neurodevelopment. Such
environmental conditions. liability also includes unique, early susceptibility to
Here, however, in the search for an epigenetic substrate social environmental exposures, and Nelson et al. (Nel-
of differential susceptibility, caution is once again son, 2014) have shown, in a random-assignment trial of
essential. The putative plasticity genes examined in foster care placements for children in Romanian orphan-
Belsky and Beaver (Belsky & Beaver, 2011) are a salad of ages, how neurobiological and developmental outcomes
diverse genomic locations, serving diverse neurobiolog- are dramatically improved when placements occur prior
ical functions. There are likely polygenic networks to 2 years of age (Zeanah, Gunnar, McCall, Kreppner &
involved in the production of risk for special suscep- Fox, 2011).
tibilities to social contexts; such allelic variation is The molecular substrates for the occurrence of such
unlikely to operate in a single direction or be capable critical periods their openings and closings across
of generating a conserved and substantial subpopulation developmental time are being elucidated within animal
of exceptionally susceptible phenotypes; and because models involving experimental manipulations at neuro-
sensitivity may be context-specific (see e.g. Obradovic, nal and molecular levels. Takesian and Hensch (Takesian
Bush & Boyce, 2011), different groups of genes may be & Hensch, 2013), for example, have shown how molec-
implicated in different categories of context sensitivity. ular triggers and brakes initiate and constrain plas-
ticity in the brain over time. The persistent loss of visual
acuity in amblyopia fails to occur, for example, when
Critical periodicity, developmental time and the cortical inhibitory circuitry, formed by GABAergic
epigenome interneurons (which use the inhibitory neurotransmitter
gamma-amino butyric acid), is compromised. Critical
Developmental time is not uniform in its potency and period onset appears guided and timed by the matura-
influence. Rather, the effects of experience change tion of excitatoryinhibitory (EI) circuit balance. The
dynamically across the lifespan, especially in the early closure of such plasticity is regulated by structural and
years, as critical and sensitive periods open and close. functional molecular brakes, such as perineuronal nets
Critical periods are those in which the presence or and the expression of proteins such as Lynx1, which
absence of important experiences or exposures result in dampen plasticity by binding to and reducing the
irreversible change in brain circuitry, while sensitive function of nicotinic acetylcholine receptors on specific
periods are developmental intervals in which the brain is inhibitory interneurons. Perineuronal net degradation or
especially responsive to such experience (Fox, Levitt & genetic deletion of Lynx1 reopens a period of plasticity
Nelson, 2010). Both involve experience-dependent to restore visual acuity to adult amyblyopic animals

2014 John Wiley & Sons Ltd

14 W. Thomas Boyce and Michael S. Kobor

(Pizzorusso, Medini, Landi, Baldini, Berardi et al., 2006; atric disorders in the generation that follows, despite an
Morishita, Miwa, Heintz & Hensch, 2010). Such findings absence of actual exposure (Yehuda, Halligan & Bierer,
have led to a fundamental shift in thinking about brain 2001). At the level of stress physiology, prenatal expo-
plasticity, from an assumption that plasticity arises sures to stressors in both animal and human mothers
during sharply defined critical periods, like square waves have been associated with differences in autonomic and
in the course of development, to a new understanding adrenocortical responses to challenge in offspring,
that the brain is intrinsically, obligatorily plastic and that though such effects substantially vary with exposure
normal development requires a timed, molecular sup- timing, offspring sex and other factors (Matthews &
pression of that plasticity. Phillips, 2012). These recapitulations of disorders or risk
Much of the molecular machinery underlying critical factors within new generations have been presumed in
period onset and offset is epigenetic in origin (Fagiolini, humans and demonstrated in mice to be mediated by
Jensen & Champagne, 2009). Epigenetic control of gene epigenetic marks transmitted to offspring (e.g. Saavedra-
expression guides, for example, the differentiation of Rodriguez & Feig, 2013).
neurons into unique neuronal subsets, the guidance of Such cross-generational propagation of epigenetic
axon growth, and the radial organization of brain marks is believed to occur via one of two possible
development (Fox et al., 2010). Brain circuitry responds mechanisms: either through behavioral and social trans-
to environmental events by way of DNA methylation fer of such marks to the subsequent generation or
and post-transcriptional histone modifications (Takesian through germline transmission. Though the latter is far
& Hensch, 2013). The closure of the critical period for more difficult to prove, there is evidence for both, at least
ocular dominance acquisition, for example, involves the in animals. The rat experiments of Meaney and col-
down-regulation of vision-dependent histone acetylation leagues is the now classical example of behavioral/social
and phosphorylation. Epigenetic factors also regulate transmission of marks. In that model, abundant mater-
the expression of GAD67, the gene coding for GABA nal care, indexed by high levels of maternal licking and
inhibitory neurotransmitter, and various pharmacologi- grooming of pups, changes DNA methylation through-
cal agents targeting epigenetic processes drugs such as out the genome (McGowan, Suderman, Sasaki, Huang,
valproate, an HDAC inhibitor can shift the timing of Hallett et al., 2010) and at a transcription factor binding
critical periods. Valproate has been shown, for example, site in the hippocampal GR promoter (Weaver, La
to reopen the critical window for the acquisition of Plante, Weaver, Parent, Sharma et al., 2001). The
absolute pitch (Gervain, Vines, Chen, Seo, Hensch et al., maternal care-associated hypomethylation of the GR
2013), and EI circuitry imbalance and critical period promoter increases expression of the receptor protein,
timing errors have been recognized within mouse models thereby down-regulating HPA reactivity and producing
of autism spectrum disorder (Gogolla, Leblanc, Quast, stress resilience in offspring. These effects and the altered
Sudhof, Fagiolini et al., 2009). These discoveries methylation status are passed on to subsequent genera-
together reveal a complex critical periodicity within tions through the mechanism of maternal behavior, the
development both adaptive and maladaptive that is down-regulated reactivity predisposing the new genera-
likely initiated, guided and curtailed by epigenetic, tion of mothers to their own high levels of pup licking
molecular events affecting the neuroregulatory genes and grooming. In presumed, human analogs of the same
that govern brain development. behavioral/social transmission of acquired epigenetic
marks, studies have documented differences in stress
reactivity among the offspring of mothers affected by the
Transgenerational inheritance of epigenetic Dutch Hunger Winter and alterations in cortisol expres-
marks sion in children who were in utero during the 2001 World
Trade Center collapse (Matthews & Phillips, 2012).
There is now substantial evidence in both humans and There is also emerging evidence, however, that some
animals that adverse physical and psychological expo- alleles in germ cells can show meiotic heritability,
sures in one generation can be replicated among or allowing epigenetic marks to be directly propagated
transmitted into subsequent generations and that such through gametes and challenging the doctrine that DNA
transgenerational diffusion can alter risks for psycho- sequences are the exclusive, heritable arbiter of pheno-
pathology and maladaptive behavior in offspring type (Whitelaw & Whitelaw, 2006). Most DNA methyl-
(Franklin, Russig, Weiss, Graff, Linder et al., 2010). ation, as noted above, is erased during early
Epidemiologic studies in human populations sustaining embryogenesis, ensuring the pluripotency of the embryo,
historic generational adversities or existential threats, for but recent exceptions have been reported that allow the
example, have documented the elevated rates of psychi- intergenerational transmission of methylated DNA,

2014 John Wiley & Sons Ltd

Geneenvironment interplay 15

post-translational histone modifications, and small, non- molecular mechanisms that may underlie G9E
coding RNAs (Bohacek et al., 2013). Reprogramming in interaction effects on developmental and mental health
the germline related to environmental exposures has now outcomes. Although such findings are clearly provisional
been demonstrated in plants, fruit flies and mammals at present, epigenetic mechanisms constitute a genuine,
(Franklin et al., 2010). These findings indicate clear, but physical nexus between environments and genes, between
still developing evidence for both forms of transgener- nurture and nature, between the exterior and interior
ational inheritance of epigenetic risk, with the strongest determinants of human development and psychological
findings in animal models. well being.
Other possible epigenetic mechanisms for the convey- Beyond this principal conclusion, derived from a
ance of parental adversity from one generation to the rapidly emerging field on the frontier of human biology,
next (in human and animal models) are the dysregulation this paper has highlighted several specific observations
of stress response pathways by the developing placenta that together trace a growing, if yet hazy, perimeter of
(Bale, Baram, Brown, Goldstein, Insel et al., 2010; Bale, developmental epigenetics. Among these are the follow-
2011), the transmission of small ncRNAs in the sperm of ing:
a traumatized father (Gapp, Jawaid, Sarkies, Bohacek, The epigenome is a structural overlay of genomic,
Pelczar et al., 2014), and an epigenetically mediated chemical markings, which target DNA and histone
transfer of fear-conditioning by an olfactory signal (Dias proteins, alter the physical packaging of chromatin,
& Ressler, 2014). Recent work by Yehuda and colleagues and regulate the expression of genes, without altering
(Yehuda, Daskalakis, Lehrner, Desarnaud, Bader et al., the nucleotide sequence itself; epigenetic processes, in
2014), moreover, has demonstrated, in the adult off- embryologic development, are the mechanisms for an
spring of Holocaust survivors, differential methylation of enduring differentiation of cells into histologically
the exon 1F promoter of the GR gene (the human distinctive cell lines.
homologue of exon 17 in the rat) according to the PTSD Experiences and environmental exposures, especially
status of the mother and father. Some or all of these those in early life, can also result in the placement or
epigenetic mechanisms may be eventually implicated in removal of epigenetic marks, thereby regulating the
human transgenerational inheritance as the science of neurodevelopment that underlies learning, behavior
such transmission grows and matures. and risks for compromised mental health.
These contrasting ontogenetic roles cell line differ-
entiation and a recording of contextual experience
Conclusions thus result in an epigenetic paradox, in which the
same epigenome becomes the origin of both the
The research summarized here arguably presages a longitudinal stability of the bodys cellular structure
period of remarkable progress in understanding the and its moment-to-moment plasticity in response to
convergence of genetic and environmental variation in environmental events.
the genesis of typical and adversity-related, atypical The G9E interactions increasingly documented
development. The epigenetic processes that appear to within the developmental and mental health litera-
mediate such convergence comprise a broad, complex set ture are likely mediated, in part, through epigenetic
of molecular points of connection between the experi- events that allow gene effects to be contingent upon
ences of early life and the proclivities, capacities and experience and experiential influences to be condi-
risks encoded within the individual human genome. The tional upon allelic variation; the epigenome thus
adversities inherent within environments of poverty, serves as a buffer to the extremes of both genetic and
neglect and trauma are transduced into molecular events environmental variation.
controlling the expression of neuroregulatory genes, Within virtually every contemporary society, the
which in turn guide brain development, calibrate stress developmental and health effects of early exposures
reactivity, and influence lifelong risks of psychopathol- to adversity and stress are socioeconomically parti-
ogy and other morbidities. Similarly, positive early tioned, with children from the lower ranks of social
environments of nurturance, care and stability provide class sustaining greater and more severe threats to
an anticipatory programming of many of the same genes, normative development; many of these pervasive SES
thereby diminishing mental health risks, optimizing influences on adversity-related, maladaptive out-
neurodevelopmental preparation for learning, and ensur- comes are almost certainly epigenetically mediated.
ing normative socioemotional development. Increas- In addition to the well-documented main effects of
ingly, the epigenetic marks and modifications childhood stress on health and development, there
controlling gene expression are being recognized as are readily observable individual differences in the

2014 John Wiley & Sons Ltd

16 W. Thomas Boyce and Michael S. Kobor

consequences of such exposures. A relatively small Prvulovic, 2013) have suggested, for example, that DNA
subset of children appear differentially susceptible to methylation profiles among genes with known linkages
the character of their rearing environments, sustain- to psychopathology could be useful as epigenetic bio-
ing exceptionally poor outcomes in contexts of markers of Major Depressive Disorder. Finally, the
adversity and threat, but unusually positive outcomes inclusion of epigenetic measures in studies of how genes
in nurturant, supportive settings; there is evidence and environments work together to affect developmental
that this differential susceptibility to social environ- outcomes may also provide insights into why either or
mental influence is also epigenetically mediated. both such etiologic factors may appear to have effect
Environmental influences are modulated by critical sizes that are modest in magnitude (Kraemer, 2012). In
periods in development, when neurobiological cir- the setting of a true, cross-over, G9E interaction, for
cuitry is especially responsive to experience and example, the association of both genetic and environ-
plasticity is most accessible; the opening and closing mental measures may appear vanishingly small, while the
of critical and sensitive periods are regulated by effect of the interaction term is substantial and unat-
epigenetic events that guide the maturation of excit- tributable to chance. In such a case, measurement of
atory and inhibitory neural circuitry and the expres- chromatin modifications within the subsets of partici-
sion of molecular brakes that reverse the brains pants defined by interaction categories may substantiate
inherent plasticity. an inference that certain combinations of genes and
Epigenetic processes are also the candidate mecha- environments are linkable to differences in gene expres-
nisms for the transmission of risk and disorder from sion and thus phenotype. Epigenetic differences might be
one generation to the next; such transmission appears legitimately viewed, in such a setting, as an endopheno-
to occur either through transgenerational replications typic precursor of symptoms or disorder (Flint &
of behavioral risk and protective factors or through Munafo, 2007).
germ line transfers of epigenetic marks. The reflexive exhilaration that attends the opening of
whole new arenas of scientific investigation has been an
The implications of these emerging findings are legion
especially visible (and contagious) aspect of the epige-
within the domains of developmental science and med-
netic bubble currently driving the bioscientific econ-
icine. First, the advent of pharmacologic agents capable
omy. The epidemic proliferation of epigenetic
of altering basic epigenetic processes holds the promise
publications and reports, with which this review began,
of entirely new classes of medication for psychiatric
disorders (Menke, Klengel & Binder, 2012) and devel- is testimony enough to the excitement with which the
industry of science has greeted this field. The euphoria
opmental disabilities (Miyake, Hirasawa, Koide & Ku-
should be tempered, however, by the limiting realities
bota, 2012). HDAC inhibitors, for example, have been
that also characterize the rapid appearance of new fields
shown to mimic the effects of antidepressants in socially
and new technologies. As recently summarized by Mill
stressed mice (Covington, Maze, LaPlant, Vialou, Ohni-
and Heijmans ((Mill & Heijmans, 2013), the emerging
shi et al., 2009), and epigenetically active agents have
field of epigenetic epidemiology is a veritable minefield of
begun to be used in the treatment of human mental
opportunities for statistical error, misinterpretation of
disorders and other complex diseases (Ptak & Petronis,
2008). Epigenetic processes are accessible, as well, by way findings, and fallacious inference. Among the pitfalls
that characterize this nascent research territory are these:
of dietary interventions that can provide or alter
bioactive components of food or the microbiota mistaking the substantial, random variation in epi-
(Shenderov & Midtvedt, 2014). Second, measured epi- genetic marks for systematic differences attributable
genetic variation might usefully serve as a proxy for to environmental exposures;
unmeasured environmental influences, as in the report by conflating peripheral and central tissue methylomes,
van IJzendoorn and colleagues (van IJzendoorn, Cas- by assuming that chromatin marks found in periph-
pers, Bakermans-Kranenburg, Beach & Philibert, 2010), eral tissues will also be present in central tissues, such
where differential methylation, a stand-in for environ- as brain;
mental adversity, interacted with 5HTTLPR allelic failing to recognize and adjust for variables, such as
variation in the prediction of unresolved loss or trauma. age, development and gender, that may confound
Third, epigenetic biomarkers might also usefully serve as associations between exposures and epigenetic
closely proximal indicators of success or failure within marks;
new therapeutic or intervention strategies, as well as employing admittedly practical and useful tools, such
signals of the causal pathways by which such successes as the Illumina 450K Human Methylation array,
are achieved. Schneider and Prvulovic (Schneider & which targets less than 2% of the CpG sites in the

2014 John Wiley & Sons Ltd

Geneenvironment interplay 17

human genome and may over-focus on promoter verifiable chromatin modifications, actually explain
regions and CpG islands; epidemiologically observed transactions between genes
using either too inclusive or too conservative data and environments. Developmental science is thus poised
analytic strategies, with their respective effects on on the cusp of a truly new molecular account at the
estimates of Type I error; very synapse between genes and contexts of the
focusing on DNA methylation because of its acces- enormous and consequential human differences in
sibility and ease of measurement, to the exclusion of development and mental health. There is no shortage
either other chromatin marks with more reliable of epigenetic terra incognita yet to explore, and though
linkages to gene expression or novel marks with as each generation of biologists may be predisposed by its
yet unknown functional roles; own genes and contexts to the same lofty and hopeful
ignoring DNA sequence polymorphisms that can claims for its science, it is a brilliant and formidable time,
alter the likelihood or effects of chromatin modifica- even now, to be exploring.
tions; and
inadequately attending to the biological conse-
quences of marks and their possible clustering within Acknowledgements
functional groups identifiable through genome anno-
tations (see e.g. Bock, 2012). The research upon which this report is based was
supported by the National Institute of Mental Health
Further, not all differential gene expression is medi-
(grant award R24MH081797), the MacArthur Founda-
ated by epigenetic modifications, as shown by Alt et al.
tion Research Network on Psychopathology and Devel-
(Alt et al., 2010), where the expression of GR in different
opment, the Sunny Hill Health Centre/British Columbia
areas of post-mortem human brain was attributable to
Leadership Chair in Child Development at the Univer-
differential expression of the NGFI-A transcription
sity of British Columbia, and the Child and Brain
factor, rather than DNA methylation. Not only do
Development Program of the Canadian Institute for
epigenetic processes not explain all differential gene
Advanced Research.
expression, but even for those questions where it should
provide tractable answers, it has sometimes failed. For
example, although a variety of studies have promisingly
documented a longitudinal divergence in epigenetic References
marks among monozygotic twins, suggesting a possible
epigenetic substrate for twin discordance (see e.g. Fraga, Adkins, R.M., Krushkal, J., Tylavsky, F.A., & Thomas, F.
Ballestar, Paz, Ropero, Setien et al., 2005), a recent (2011). Racial differences in gene-specific DNA methylation
paper by Baranzini and colleagues (Baranzini, Mudge, levels are present at birth. Birth Defects Research: Part A,
van Velkinburgh, Khankhanian, Khrebtukova et al., Clinical and Molecular Teratology, 91 (8), 728736.
2010) reported no epigenetic or transcriptome differ- Allis, C.D., Jenuwein, T., Reinberg, D., & Caparros, M.-L.
(2007). Epigenetics. Cold Spring Harbor, NY: Cold Spring
ences in a small sample of monozygotic twins discordant
Harbor Laboratory Press.
for multiple sclerosis. Such findings, even on a small Alt, S.R., Turner, J.D., Klok, M.D., Meijer, O.C., Lakke, E.A.
scale, urge caution and restraint in the interpretation et al. (2010). Differential expression of glucocorticoid recep-
(and celebration) of the epigenetic juggernaut. There tor transcripts in major depressive disorder is not epigenet-
remains much unknown about how genes and environ- ically programmed. Psychoneuroendocrinology, 35 (4), 544
ments converge in effects, how chromatin modification is 556.
linked to gene expression, and how the developmental Babineau, V., Gordon Green, C., Jolicoeur-Martineau, A.,
transcriptome determines phenotypic ends. Epigenetic Minde, K., Sassi, R. et al. (2014). Prenatal depression and
biology is thus a field filled with an early harvest of 5-HTTLPR interact to predict dysregulation from 3 to 36
appealing but preliminary findings a field that should months: a differential susceptibility model. Journal of Child
be regarded, at present, as both fertile and fetal. Psychology and Psychiatry, and Allied Disciplines. doi: 10.
Such scientific realities notwithstanding, there is much
Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H.
to admire in the heuristic productivity of early epigenetic (2011). Differential susceptibility to rearing environment
research in human populations. In arguably no more depending on dopamine-related genes: new evidence and a
than a decade, the field has progressed from a tentative meta-analysis. Development and Psychopathology, 23 (1), 39
documentation of G9E interactions, to theoretical 52.
expositions on how such interactions might occur, to Bale, T.L. (2011). Sex differences in prenatal epigenetic
real evidence that molecular mechanisms, involving programming of stress pathways. Stress, 14 (4), 348356.

2014 John Wiley & Sons Ltd

18 W. Thomas Boyce and Michael S. Kobor

Bale, T.L., Baram, T.Z., Brown, A.S., Goldstein, J.M., Insel, position in adult DNA methylation. International Journal of
T.R. et al. (2010). Early life programming and neurode- Epidemiology, 41 (1), 6274.
velopmental disorders. Biological Psychiatry., 68 (4), 314 Boyce, W.T., Chesney, M., Alkon-Leonard, A., Tschann, J.,
319. Adams, S. et al. (1995). Psychobiologic reactivity to stress
Baranzini, S.E., Mudge, J., van Velkinburgh, J.C., Khankha- and childhood respiratory illnesses: results of two prospective
nian, P., Khrebtukova, I. et al. (2010). Genome, epigenome studies. Psychosomatic Medicine, 57, 411422.
and RNA sequences of monozygotic twins discordant for Boyce, W.T., & Ellis, B.J. (2005). Biological sensitivity to
multiple sclerosis. Nature, 464 (7293), 13511356. context: I. An evolutionary-developmental theory of the
Barr, C.S., Newman, T.K., Lindell, S., Shannon, C., Champ- origins and functions of stress reactivity. Development and
oux, M. et al. (2004a). Interaction between serotonin Psychopathology, 17 (2), 271301.
transporter gene variation and rearing condition in alcohol Boyce, W.T., Sokolowski, M.B., & Robinson, G.E. (2012).
preference and consumption in female primates. Archives of Toward a new biology of social adversity. Proceedings of the
General Psychiatry, 61 (11), 11461152. National Academy of Sciences, USA, 109 (Suppl. 2), 17143
Barr, C.S., Newman, T.K., Shannon, C., Parker, C., Dvoskin, 17148.
R.L. et al. (2004b). Rearing condition and rh5-HTTLPR Brett, Z.H., Sheridan, M.A., Jones, E.G., Esteves, K.C., Fox,
interact to influence limbic-hypothalamic-pituitary-adrenal N.A. et al. (2014). Association of two ERBB3SNP genotypes
axis response to stress in infant macaques. Biological with sensitivity to caregiving context. Boston, MA: Poster
Psychiatry, 55 (7), 733738. presented at the International Society for Psychiatry Genet-
Beach, S.R., Brody, G.H., Lei, M.K., Kim, S., Cui, J. et al. ics.
(2014). Is serotonin transporter genotype associated with Bridi, M., & Abel, T. (2013). Histone modifications in the
epigenetic susceptibility or vulnerability? Examination of the nervous system and neuropsychiatric disorders. In J.D.
impact of socioeconomic status risk on African American Sweatt, M. Meaney, E.J. Nestler & S. Akbarian (Eds.),
youth. Development and Psychopathology, 26 (2), 289304. Epigenetic regulation in the nervous system: Basic mechanisms
Bell, J.T., Pai, A.A., Pickrell, J.K., Gaffney, D.J., Pique-Regi, and clinical impact (pp. 3567). London: Elsevier.
R., et al. (2011). DNA methylation patterns associate with Broekman, B.F., Chan, Y.H., Goh, L., Fung, D., Gluckman,
genetic and gene expression variation in HapMap cell lines. P.D. et al. (2011). Influence of birth weight on internalizing
Genome Biology, 12 (1), R10. traits modulated by serotonergic genes. Pediatrics, 128 (5),
Belsky, J. (2005). Differential susceptibility to rearing influence: e1250e1258.
an evolutionary hypothesis and some evidence. In B.J. Ellis & Burns, J., Svetec, N., Rowe, L., Mery, F., Dolan, M. et al.
D.F. Bjorklund (Eds.), Origins of the social mind: Evolution- (2012). Geneenvironment interplay in Drosophila melanog-
ary psychology and child development (pp. 139163). New aster: chronic food deprivation in early life affects adult
York: Guilford. exploratory and fitness traits. Proceedings of the National
Belsky, J., & Beaver, K.M. (2011). Cumulative-genetic plasticity, Academy of Sciences, USA, 109 (Suppl. 2), 1723917244.
parenting and adolescent self-regulation. Journal of Child Bush, N., Guendelman, M., Adler, N., & Boyce, W.T. (2014,
Psychology and Psychiatry, 52 (5), 619626. submitted). BDNF allelic variants moderate social disparities
Berko, E.R., Suzuki, M., Beren, F., Lemetre, C., Alaimo, C.M. in childrens basal cortisol expression.
et al. (2014). Mosaic epigenetic dysregulation of ectodermal Caspi, A., Hariri, A.R., Holmes, A., Uher, R., & Moffitt, T.E.
cells in autism spectrum disorder. PLoS Genetics, 10 (5), (2010). Genetic sensitivity to the environment: the case of the
e1004402. serotonin transporter gene and its implications for studying
Bird, A. (2007). Perceptions of epigenetics. Nature, 447 (7143), complex diseases and traits. American Journal of Psychiatry,
396398. 167, 509527.
Bock, C. (2012). Analysing and interpreting DNA methylation Caspi, A., McClay, J., Moffitt, T.E., Mill, J., Martin, J. et al.
data. Nature Reviews Genetics, 13 (10), 705719. (2002). Role of genotype in the cycle of violence in
Bogdan, R., Agrawal, A., Gaffrey, M.S., Tillman, R., & Luby, maltreated children. Science, 297 (5582), 851854.
J.L. (2014). Serotonin transporter-linked polymorphic region Caspi, A., Sugden, K., Moffitt, T.E., Taylor, A., Craig, I.W. et
(5-HTTLPR) genotype and stressful life events interact to al. (2003). Influence of life stress on depression: moderation
predict preschool-onset depression: a replication and devel- by a polymorphism in the 5-HTT gene. Science, 301 (5631),
opmental extension. Journal of Child Psychology and Psy- 386389.
chiatry, 55 (5), 448457. Champagne, F.A. (2010). Early adversity and developmental
Bohacek, J., Gapp, K., Saab, B.J., & Mansuy, I.M. (2013). outcomes: interaction between genetics, epigenetics, and
Transgenerational epigenetic effects on brain functions. social experiences across the life span. Perspectives on
Biological Psychiatry, 73 (4), 313320. Psychological Science, 5 (5), 564574.
Bojang, P. Jr., & Ramos, K.S. (2014). The promise and failures Cole, S.W., Conti, G., Arevalo, J.M., Ruggiero, A.M., Heck-
of epigenetic therapies for cancer treatment. Cancer Treat- man, J.J. et al. (2012). Transcriptional modulation of the
ment Reviews, 40 (1), 153169. doi:10.1016/j.ctrv.2013.05.009 developing immune system by early life social adversity.
Borghol, N., Suderman, M., McArdle, W., Racine, A., Hallett, Proceedings of the National Academy of Sciences, USA, 109
M. et al. (2012). Associations with early life socio-economic (50), 2057820583.

2014 John Wiley & Sons Ltd

Geneenvironment interplay 19

Covington, H.E. 3rd, Maze, I., LaPlant, Q.C., Vialou, V.F., monozygotic twins. Proceedings of the National Academy of
Ohnishi, Y.N. et al. (2009). Antidepressant actions of histone Sciences, USA, 102 (30), 1060410609.
deacetylase inhibitors. Journal of Neuroscience, 29 (37), Franklin, T.B., Russig, H., Weiss, I.C., Graff, J., Linder, N. et al.
1145111460. (2010). Epigenetic transmission of the impact of early stress
Danese, A., Pariante, C.M., Caspi, A., Taylor, A. & Poulton, R. across generations. Biological Psychiatry, 68 (5), 408415.
(2007). Childhood maltreatment predicts adult inflammation Fraser, H.B., Lam, L.L., Neumann, S.M., & Kobor, M.S.
in a life-course study. Proceedings of the National Academy of (2012). Population-specificity of human DNA methylation.
Sciences, USA, 104 (4), 13191324. Genome Biology, 13 (2), R8.
Davies, M.N., Volta, M., Pidsley, R., Lunnon, K., Dixit, A. et Gapp, K., Jawaid, A., Sarkies, P., Bohacek, J., Pelczar, P. et al.
al. (2012). Functional annotation of the human brain (2014). Implication of sperm RNAs in transgenerational
methylome identifies tissue-specific epigenetic variation inheritance of the effects of early trauma in mice. Nature
across brain and blood. Genome Biology, 13 (6), R43. Neuroscience, 17, 667669.
Dekker, A.D., De Deyn, P.P., & Rots, M.G. (2014). Epigenetics: Gervain, J., Vines, B.W., Chen, L.M., Seo, R.J., Hensch, T.K. et
the neglected key to minimize learning and memory deficits al. (2013). Valproate reopens critical-period learning of
in Down syndrome. Neuroscience & Biobehavioral Reviews, absolute pitch. Frontiers in Systems Neuroscience, 7, 102.
45C, 7284. Gibbs, J.R., van der Brug, M.P., Hernandez, D.G., Traynor,
Dias, B.G., & Ressler, K.J. (2014). Parental olfactory experience B.J., Nalls, M.A. et al. (2010). Abundant quantitative trait
influences behavior and neural structure in subsequent loci exist for DNA methylation and gene expression in
generations. Nature Neuroscience, 17 (1), 8996. human brain. PLoS Genetics, 6 (5), e1000952.
Drury, S.S., Theall, K.P., Smyke, A.T., Keats, B.J., Egger, H.L. Gilbert, S.F., & Epel, D. (2009). Ecological developmental
et al. (2010). Modification of depression by COMT biology: Integrating epigenetics, medicine, and evolution.
val158met polymorphism in children exposed to early severe Sunderland, MA: Sinauer Associates.
psychosocial deprivation. Child Abuse & Neglect, 34 (6), 387 Globisch, D., Muenzel, M., Mueller, M., Michalakis, S.,
395. Wagner, M. et al. (2010). Tissue distribution of 5-hydrox-
Duncan, L.E., & Keller, M.C. (2011). A critical review of the ymethylcytosine and search for active demethylation inter-
first 10 years of candidate gene-by-environment interaction mediates. PLoS ONE, 5 (12), e15367.
research in psychiatry. American Journal of Psychiatry, 168 Gogolla, N., Leblanc, J.J., Quast, K.B., Sudhof, T.C., Fagiolini,
(10), 10411049. M. et al. (2009). Common circuit defect of excitatory-inhib-
Ellis, B.J., & Bjorklund, D.F. (2012). Beyond mental health: an itory balance in mouse models of autism. Journal of
evolutionary analysis of development under risky and Neurodevelopmental Disorders, 1 (2), 172181.
supportive environmental conditions: an introduction to Gratten, J., Wray, N.R., Keller, M.C., & Visscher, P.M. (2014).
the special section. Developmental Psychology, 48 (3), 591 Large-scale genomics unveils the genetic architecture of
597. psychiatric disorders. Nature Neuroscience, 17 (6), 782790.
Ellis, B.J., Boyce, W.T., Belsky, J., Bakermans-Kranenburg, Gutierrez-Arcelus, M., Lappalainen, T., Montgomery, S.B.,
M.J., & van Ijzendoorn, M.H. (2011). Differential suscepti- Buil, A., & Ongen, H. et al. (2013). Passive and active DNA
bility to the environment: an evolutionary-neurodevelop- methylation and the interplay with genetic variation in gene
mental theory. Development and Psychopathology, 23 (1), 7 regulation. eLife, 2, e00523.
28. Heijmans, B.T., Tobi, E.W., Stein, A.D., Putter, H., Blauw, G.J.
Ellis, B.J., Essex, M.J., & Boyce, W.T. (2005). Biological et al. (2008). Persistent epigenetic differences associated with
sensitivity to context: II. Empirical explorations of an prenatal exposure to famine in humans. Proceedings of the
evolutionary-developmental hypothesis. Development and National Academy of Sciences, USA, 105 (44), 1704617049.
Psychopathology, 17 (2), 303328. Hertzman, C., & Boyce, W.T. (2010). How experience gets
Essex, M.J., Boyce, W.T., Hertzman, C., Lam, L.L., Armstrong, under the skin to create gradients in developmental health.
J.M. et al. (2013). Epigenetic vestiges of early developmental Annual Review of Public Health, 31, 329347.
adversity: childhood stress exposure and DNA methylation Heyn, H., Moran, S., Hernando-Herraez, I., Sayols, S., Gomez,
in adolescence. Child Development, 84 (1), 5875. A. et al. (2013). DNA methylation contributes to natural
Fagiolini, M., Jensen, C.L., & Champagne, F.A. (2009). human variation. Genome Research, 23 (9), 13631372.
Epigenetic influences on brain development and plasticity. Hodes, G.E. (2013). Sex, stress, and epigenetics: regulation of
Current Opinion in Neurobiology, 19 (2), 207212. behavior in animal models of mood disorders. Biology of Sex
Flint, J., & Munafo, M.R. (2007). The endophenotype concept Differences, 4 (1), 1.
in psychiatric genetics. Psychological Medicine, 37 (2), 163 Horvath, S., Zhang, Y., Langfelder, P., Kahn, R.S., Boks, M.P.
180. et al. (2012). Aging effects on DNA methylation modules in
Fox, S.E., Levitt, P., & Nelson, C.A. (2010). How the timing human brain and blood tissue. Genome Biology, 13 (10), R97.
and quality of early experiences influence the development of Houseman, E.A., Accomando, W.P., Koestler, D.C., Christen-
brain architecture. Child Development, 81 (1), 2840. sen, B.C., Marsit, C.J. et al. (2012). DNA methylation arrays
Fraga, M.F., Ballestar, E., Paz, M.F., Ropero, S., Setien, F., et as surrogate measures of cell mixture distribution. BMC
al. (2005). Epigenetic differences arise during the lifetime of Bioinformatics, 13, 86.

2014 John Wiley & Sons Ltd

20 W. Thomas Boyce and Michael S. Kobor

Illingworth, R.S., & Bird, A.P. (2009). CpG islands a rough tissue than blood for epigenome-wide association studies.
guide. FEBS Letters, 583 (11), 17131720. Epigenetics, 8 (4), 445454.
Jaffe, A.E., & Irizarry, R.A. (2014). Accounting for cellular Lubin, F.D., Roth, T.L., & Sweatt, J.D. (2008). Epigenetic
heterogeneity is critical in epigenome-wide association stud- regulation of BDNF gene transcription in the consolidation
ies. Genome Biology, 15 (2), R31. of fear memory. Journal of Neuroscience, 28 (42), 10576
Jones, M.J., Fejes, A.P., & Kobor, M.S. (2013). DNA methyl- 10586.
ation, genotype and gene expression: who is driving and who McGowan, P.O., Sasaki, A., DAlessio, A.C., Dymov, S.,
is along for the ride? Genome Biology, 14 (7), 126. Labonte, B. et al. (2009). Epigenetic regulation of the
Jones, P.A. (1999). The DNA methylation paradox. Trends in glucocorticoid receptor in human brain associates with
Genetics, 15 (1), 3437. childhood abuse. Nature Neuroscience, 12 (3), 342348.
Karg, K., Burmeister, M., Shedden, K., & Sen, S. (2011). The McGowan, P.O., Suderman, M., Sasaki, A., Huang, T.C.,
serotonin transporter promoter variant (5-HTTLPR), stress, Hallett, M. et al. (2010). Broad epigenetic signature of
and depression meta-analysis revisited: evidence of genetic maternal care in the brain of adult rats. PLoS One, 6 (2),
moderation. Archives in General Psychiatry, 68 (5), 444454. e14739.
Kass, S.U., Landsberger, N., & Wolffe, A.P. (1997). DNA Marigorta, U.M., & Gibson, G. (2014). A simulation study of
methylation directs a time-dependent repression of transcrip- gene-by-environment interactions in GWAS implies ample
tion initiation. Current Biology, 7 (3), 157165. hidden effects. Frontiers in Genetics, 5, 225.
Kinnally, E.L., Feinberg, C., Kim, D., Ferguson, K., Leibel, R. Masten, A.S. (2014). Global perspectives on resilience in
et al. (2011). DNA methylation as a risk factor in the effects children and youth. Child Development, 85 (1), 620.
of early life stress. Brain, Behavior, and Immunity, 25 (8), Matthews, S.G., & Phillips, D.I. (2012). Transgenerational
15481553. inheritance of stress pathology. Experimental Neurology, 233
Klengel, T., Mehta, D., Anacker, C., Rex-Haffner, M., Pruess- (1), 95101.
ner, J.C. et al. (2013). Allele-specific FKBP5 DNA demethy- Mattick, J.S. (2010). RNA as the substrate for epigenome
lation mediates genechildhood trauma interactions. Nature environment interactions: RNA guidance of epigenetic
Neuroscience, 16 (1), 3341. processes and the expansion of RNA editing in animals
Klengel, T., Pape, J., Binder, E.B., & Mehta, D. (2014). The role underpins development, phenotypic plasticity, learning, and
of DNA methylation in stress-related psychiatric disorders. cognition. Bioessays, 32 (7), 548552.
Neuropharmacology, 80, 115132. Meaney, M.J. (2001). Maternal care, gene expression, and the
Kofink, D., Boks, M.P., Timmers, H.T., & Kas, M.J. (2013). transmission of individual differences in stress reactivity
Epigenetic dynamics in psychiatric disorders: environmental across generations. Annual Review of Neuroscience, 24, 1161
programming of neurodevelopmental processes. Neurosci- 1192.
ence and Biobehavioral Reviews, 37, 831845. Meaney, M.J. (2010). Epigenetics and the biological definition
Korosi, A., Naninck, E.F., Oomen, C.A., Schouten, M., of gene 9 environment interactions. Child Development, 81
Krugers, H. et al. (2012). Early-life stress mediated modu- (1), 4179.
lation of adult neurogenesis and behavior. Behavioural Brain Mehta, D., Klengel, T., Conneely, K.N., Smith, A.K., Altmann,
Research, 227 (2), 400409. A. et al. (2013). Childhood maltreatment is associated with
Kraemer, H.C. (2012). Determining gene moderation of distinct genomic and epigenetic profiles in posttraumatic
environmental risk factors for a mental disorder: a perfect stress disorder. Proceedings of the National Academy of
storm of methodological problems. International Journal of Sciences, USA, 110 (20), 83028307.
Methods in Psychiatric Research, 21 (3), 185194. Menke, A., Klengel, T., & Binder, E.B. (2012). Epigenetics,
Kuzawa, C.W., & Thayer, Z.M. (2011). Timescales of human depression and antidepressant treatment. Current Pharma-
adaptation: the role of epigenetic processes. Epigenomics, 3 ceutical Design, 18 (36), 58795889.
(2), 221234. Mill, J., & Heijmans, B.T. (2013). From promises to practical
Lam, L.L., Emberly, E., Fraser, H.B., Neumann, S.M., Chen, strategies in epigenetic epidemiology. Nature Reviews Genet-
E. et al. (2012). Factors underlying variable DNA methyl- ics, 14 (8), 585594.
ation in a human community cohort. Proceedings of the Miyake, K., Hirasawa, T., Koide, T., & Kubota, T. (2012).
National Academy of Sciences, USA, 109 (Suppl. 2), 17253 Epigenetics in autism and other neurodevelopmental dis-
17260. eases. Advances in Experimental Medicine and Biology, 724,
Lister, R., Mukamel, E.A., Nery, J.R., Urich, M., Puddifoot, 9198.
C.A. et al. (2013). Global epigenomic reconfiguration during Moen, E.L., Zhang, X., Mu, W., Delaney, S.M., Wing, C. et al.
mammalian brain development. Science, 341 (6146), 1237905. (2013). Genome-wide variation of cytosine modifications
Lister, R., Pelizzola, M., Dowen, R.H., Hawkins, R.D., Hon, between European and African populations and the impli-
G. et al. (2009). Human DNA methylomes at base resolution cations for complex traits. Genetics, 194 (4), 987996.
show widespread epigenomic differences. Nature, 462 (7271), Monk, C., Spicer, J., & Champagne, F.A. (2012). Linking
315322. prenatal maternal adversity to developmental outcomes in
Lowe, R., Gemma, C., Beyan, H., Hawa, M.I., Bazeos, A. et al. infants: the role of epigenetic pathways. Development and
(2013). Buccals are likely to be a more informative surrogate Psychopathology, 24 (4), 13611376.

2014 John Wiley & Sons Ltd

Geneenvironment interplay 21

Morishita, H., Miwa, J.M., Heintz, N., & Hensch, T.K. (2010). Provencal, N., Suderman, M.J., Guillemin, C., Massart, R.,
Lynx1, a cholinergic brake, limits plasticity in adult visual Ruggiero, A. et al. (2012). The signature of maternal rearing
cortex. Science, 330 (6008), 12381240. in the methylome in rhesus macaque prefrontal cortex and T
Mueller, B.R., & Bale, T.L. (2008). Sex-specific programming of cells. Journal of Neuroscience, 32 (44), 1562615642.
offspring emotionality after stress early in pregnancy. Journal Ptak, C., & Petronis, A. (2008). Epigenetics and complex
of Neuroscience, 28 (36), 90559065. disease: from etiology to new therapeutics. Annual Review of
Murphy, T.M., Mullins, N., Ryan, M., Foster, T., Kelly, C. et al. Pharmacology and Toxicology, 48, 257276.
(2013). Genetic variation in DNMT3B and increased global Qureshi, I.A., & Mehler, M.F. (2010). Genetic and epigenetic
DNA methylation is associated with suicide attempts in underpinnings of sex differences in the brain and in neuro-
psychiatric patients. Genes, Brain, and Behavior, 12 (1), 125132. logical and psychiatric disease susceptibility. Progress in
National Scientific Council on the Developing Child (2005/ Brain Research, 186, 7795.
2014). Excessive stress disrupts the architecture of the Radtke, K.M., Ruf, M., Gunter, H.M., Dohrmann, K., Schauer,
developing brain. Working Paper 3. Updated edition. M. et al. (2011). Transgenerational impact of intimate partner
Retrieved from: violence on methylation in the promoter of the glucocorticoid
Naumova, O.Y., Lee, M., Koposov, R., Szyf, M., Dozier, M. et receptor. Translational Psychiatry, 1, e21.
al. (2012). Differential patterns of whole-genome DNA Rietveld, C.A., Medland, S.E., Derringer, J., Yang, J., Esko, T.
methylation in institutionalized children and children raised et al. (2013). GWAS of 126,559 individuals identifies genetic
by their biological parents. Development and Psychopathol- variants associated with educational attainment. Science, 340
ogy, 24 (1), 143155. (6139), 14671471.
Nederhof, E., & Schmidt, M.V. (2012). Mismatch or cumulative Risch, N., Herrell, R., Lehner, T., Liang, K.Y., Eaves, L. et al.
stress: toward an integrated hypothesis of programming (2009). Interaction between the serotonin transporter gene
effects. Physiology & Behavior, 106 (5), 691700. (5-HTTLPR), stressful life events, and risk of depression: a
Nelson, C.A. (2014). Romanias abandoned children. Cam- meta-analysis. Journal of the American Medical Association,
bridge, MA: Harvard University Press. 301, 24622471.
Oberlander, T.F., Weinberg, J., Papsdorf, M., Grunau, R., Romens, S.E., Svaren, J., & Pollak, S.D. (2014, in press).
Misri, S. et al. (2008). Prenatal exposure to maternal Associations between early life stress and gene methylation in
depression, neonatal methylation of human glucocorticoid children. Child Development. doi: 10.1111/cdev.12270
receptor gene (NR3C1) and infant cortisol stress response. Roth, T.L., Lubin, F.D., Funk, A.J., & Sweatt, J.D. (2009).
Epigenetics, 3 (2), 97106. Lasting epigenetic influence of early-life adversity on the
Obradovic, J., Bush, N.R., & Boyce, W.T. (2011). The interac- BDNF gene. Biological Psychiatry, 65 (9), 760769.
tive effect of marital conflict and stress reactivity on Rothman, K.J., & Greenland, S. (2005). Causation and causal
externalizing and internalizing symptoms: the role of labo- inference in epidemiology. American Journal of Public
ratory stressors. Development and Psychopathology, 23 (1), Health, 95 (Suppl. 1), S144S150.
101114. Rutter, M. (2010). Geneenvironment interplay. Depression and
Olsson, C.A., Foley, D.L., Parkinson-Bates, M., Byrnes, G., & Anxiety, 27 (1), 14. doi:10.1002/da.20641
M. McKenzie, M. et al. (2010). Prospects for epigenetic Rutter, M. (2012). Resilience as a dynamic concept. Develop-
research within cohort studies of psychological disorder: a ment and Psychopathology, 24, 335344.
pilot investigation of a peripheral cell marker of epigenetic Rutter, M., Thapar, A., & Pickles, A. (2009). Geneenviron-
risk for depression. Biological Psychology, 83 (2), 159165. ment interactions: biologically valid pathway or artifact?
Ouellet-Morin, I., Wong, C.C., Danese, A., Pariante, C.M., Archives of General Psychiatry, 66 (12), 12871289.
Papadopoulos, A.S. et al. (2013). Increased serotonin trans- Saavedra-Rodriguez, L., & Feig, L.A. (2013). Chronic social
porter gene (SERT) DNA methylation is associated with instability induces anxiety and defective social interactions
bullying victimization and blunted cortisol response to stress across generations. Biological Psychiatry, 73 (1), 4453.
in childhood: a longitudinal study of discordant monozygotic Sasaki, A., de Vega, W.C., & McGowan, P.O. (2013). Biological
twins. Psychological Medicine, 43 (9), 18131823. embedding in mental health: an epigenomic perspective.
Petronis, A. (2010). Epigenetics as a unifying principle in the Biochemistry and Cell Biology, 91 (1), 1421.
aetiology of complex traits and diseases. Nature, 465 (7299), Saxonov, S., Berg, P., & Brutlag, D.L. (2006). A genome--
721727. wide analysis of CpG dinucleotides in the human genome
Pizzorusso, T., Medini, P., Landi, S., Baldini, S., Berardi, N. et distinguishes two distinct classes of promoters. Proceedings
al. (2006). Structural and functional recovery from early of the National Academy of Sciences, USA, 103 (5), 1412
monocular deprivation in adult rats. Proceedings of the 1417.
National Academy of Sciences, USA, 103 (22), 85178522. Schneider, B., & Prvulovic, D. (2013). Novel biomarkers in
Powell, N.D., Sloan, E.K., Bailey, M.T., Arevalo, J.M., Miller, major depression. Current Opinion in Psychiatry, 26 (1), 47
G.E. et al. (2013). Social stress up-regulates inflammatory gene 53.
expression in the leukocyte transcriptome via beta-adrenergic Shenderov, B.A., & Midtvedt, T. (2014). Epigenomic program-
induction of myelopoiesis. Proceedings of the National Acad- ing: a future way to health? Microbial Ecology in Health and
emy of Sciences, USA, 110 (41), 1657416579. Disease, 25.

2014 John Wiley & Sons Ltd

22 W. Thomas Boyce and Michael S. Kobor

Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S., & Toyokawa, S., Uddin, M., Koenen, K.C., & Galea, S. (2012).
Duman, R.S. (2002). Brain-derived neurotrophic factor How does the social environment get into the mind?
produces antidepressant effects in behavioral models of Epigenetics at the intersection of social and psychiatric
depression. Journal of Neuroscience, 22 (8), 32513261. epidemiology. Social Science & Medicine, 74 (1), 6774.
Shonkoff, J.P., Boyce, W.T., & McEwen, B.S. (2009). Neurosci- Tsankova, N., Renthal, W., Kumar, A., & Nestler, E.J. (2007).
ence, molecular biology, and the childhood roots of health Epigenetic regulation in psychiatric disorders. Nature
disparities: building a new framework for health promotion Reviews Neuroscience, 8 (5), 355367.
and disease prevention. Journal of the American Medical Tucker-Drob, E.M., Rhemtulla, M., Harden, K.P., Turkheimer,
Association, 301 (21), 22522259. E., & Fask, D. (2011). Emergence of a Gene 9 socioeco-
Shulha, H.P., Cheung, I., Guo, Y., Akbarian, S., & Weng, Z. nomic status interaction on infant mental ability between 10
(2013). Coordinated cell type-specific epigenetic remodeling months and 2 years. Psychological Science, 22 (1), 125133.
in prefrontal cortex begins before birth and continues into Tung, J., Barreiro, L.B., Johnson, Z.P., Hansen, K.D., V.
early adulthood. PLoS Genetics, 9 (4), e1003433. Michopoulos, V. et al. (2012). Social environment is associ-
Shulha, H.P., Cheung, I., Whittle, C., Wang, J., Virgil, D. et al. ated with gene regulatory variation in the rhesus macaque
(2012). Epigenetic signatures of autism: trimethylated H3K4 immune system. Proceedings of the National Academy of
landscapes in prefrontal neurons. Archives of General Psy- Sciences, USA, 109 (17), 64906495.
chiatry, 69 (3), 314324. Tyrka, A.R., Price, L.H., Marsit, C., Walters, O.C., &
Simons, R.L., Lei, M.K., Beach, S.R., Brody, G.H., Philibert, Carpenter, L.L. (2012). Childhood adversity and epigenetic
R.A. et al. (2011). Social environmental variation, plasticity modulation of the leukocyte glucocorticoid receptor: pre-
genes, and aggression: evidence for the differential susceptibil- liminary findings in healthy adults. PLoS One, 7 (1), e30148.
ity hypothesis. American Sociological Review, 76 (6), 833912. Unternaehrer, E., Luers, P., Mill, J., Dempster, E., A. H. Meyer,
Smith, Z.D., & Meissner, A. (2013). DNA methylation: roles in A.H. et al. (2012). Dynamic changes in DNA methylation of
mammalian development. Nature Reviews Genetics, 14 (3), stress-associated genes (OXTR, BDNF) after acute psycho-
204220. social stress. Translational Psychiatry, 2, e150.
Strachan, T., & Read, A. (2011). Human molecular genetics. van IJzendoorn, M.H., Caspers, K., Bakermans-Kranenburg,
New York: Garland. M.J., Beach, S.R., & Philibert, R. (2010). Methylation
Strahl, B.D., & Allis, C.D. (2000). The language of covalent matters: interaction between methylation density and sero-
histone modifications. Nature, 403 (6765), 4145. tonin transporter genotype predicts unresolved loss or
Strunk, T., Jamieson, S.E., & Burgner, D. (2013). Genetic and trauma. Biological Psychiatry, 68 (5), 405407.
epigenetic susceptibility to early life infection. Current van Winkel, M., Peeters, F., van Winkel, R., Kenis, G., Collip,
Opinion in Infectious Disease, 26 (3), 241247. D. et al. (2014). Impact of variation in the BDNF gene on
Suderman, M., McGowan, P.O., Sasaki, A., Huang, T.C., social stress sensitivity and the buffering impact of positive
Hallett, M.T. et al. (2012). Conserved epigenetic sensitivity to emotions: replication and extension of a geneenvironment
early life experience in the rat and human hippocampus. interaction. European Neuropsychopharmacology, 24 (6),
Proceedings of the National Academy of Sciences, USA, 109 930938.
(Suppl. 2), 1726617272. Weaver, I.C., Diorio, J., Seckl, J.R., Szyf, M., & Meaney, M.J.
Sweatt, J.D., Meaney, M.J., Nestler, E.J., & Akbarian, S. (2013). (2004). Early environmental regulation of hippocampal
Epigenetic regulation in the nervous system: Basic mechanisms glucocorticoid receptor gene expression: characterization of
and clinical impact. London: Elsevier. intracellular mediators and potential genomic target sites.
Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, Annals of the New York Academy of Sciences, 1024, 182212.
H. et al. (2009). Conversion of 5-methylcytosine to 5-hy- Weaver, I.C., La Plante, P., Weaver, S., Parent, A., Sharma, S. et
droxymethylcytosine in mammalian DNA by MLL partner al. (2001). Early environmental regulation of hippocampal
TET1. Science, 324 (5929), 930935. glucocorticoid receptor gene expression: characterization of
Takesian, A.E., & Hensch, T.K. (2013). Balancing plasticity/ intracellular mediators and potential genomic target sites.
stability across brain development. Progress in Brain Molecular and Cellular Endocrinology, 185 (12), 205218.
Research, 207, 334. Weaver, I.C., Meaney, M.J., & Szyf, M. (2006). Maternal care
Tarullo, A.R., & Gunnar, M.R. (2006). Child maltreatment and effects on the hippocampal transcriptome and anxiety-med-
the developing HPA axis. Hormones and Behavior, 50 (4), iated behaviors in the offspring that are reversible in
632639. adulthood. Proceedings of the National Academy of Sciences,
Teh, A.L., Pan, H., Chen, L., Ong, M.L., Dogra, S. et al. USA, 103 (9), 34803485.
(2014). The effect of genotype and in utero environment on Whitelaw, N.C., & Whitelaw, E. (2006). How lifetimes shape
inter-individual variation in neonate DNA methylomes. epigenotype within and across generations. Human Molecu-
Genome Research, 24, 10641074. lar Genetics, 15, Spec No. 2, R131R137.
Tobi, E.W., Lumey, L.H., Talens, R.P., Kremer, D., Putter, H. et Wilson, M.E., & Kinkead, B. (2008). Geneenvironment
al. (2009). DNA methylation differences after exposure to interactions, not neonatal growth hormone deficiency, time
prenatal famine are common and timing- and sex-specific. puberty in female rhesus monkeys. Biology of Reproduction,
Human Molecular Genetics, 18 (21), 40464053. 78 (4), 736743.

2014 John Wiley & Sons Ltd

Geneenvironment interplay 23

Witzmann, S.R., Turner, J.D., Meriaux, S.B., Meijer, O.C., & Zeanah, C.H., Gunnar, M.R., McCall, R.B., Kreppner, J.M.,
Muller, C.P. (2012). Epigenetic regulation of the glucocorti- & Fox, N.A. (2011). VI. Sensitive periods. Monographs of
coid receptor promoter 1(7) in adult rats. Epigenetics, 7 (11), the Society for Research in Child Development, 76 (4), 147
12901301. 162.
Wright, R.O., & Christiani, D. (2010). Geneenvironment Ziller, M.J., Gu, H., Muller, F., Donaghey, J., Tsai, L.T. et al.
interaction and childrens health and development. Current (2013). Charting a dynamic DNA methylation landscape of
Opinion in Pediatrics, 22 (2), 197201. the human genome. Nature, 500 (7463), 477481.
Wu, H., & Zhang, Y. (2014). Reversing DNA methylation: Zohsel, K., Buchmann, A.F., Blomeyer, D., Hohm, E.,
mechanisms, genomics, and biological functions. Cell, 156 Schmidt, M.H. et al. (2014). Mothers prenatal stress and
(12), 4568. their childrens antisocial outcomes a moderating role for
Yehuda, R., Daskalakis, N.P., Lehrner, A., Desarnaud, F., the Dopamine D4 Receptor (DRD4) gene. Journal of Child
Bader, H.N. et al. (2014). Influences of maternal and paternal Psychology and Psychiatry, 55 (1), 6976.
PTSD on epigenetic regulation of the glucocorticoid receptor Zou, J., Lippert, C., Heckerman, D., Aryee, M., & Listgar-
gene in Holocaust survivor offspring. American Journal of ten, J. (2014). Epigenome-wide association studies without
Psychiatry, 171 (8), 872880. the need for cell-type composition. Nature Methods, 11 (3),
Yehuda, R., Halligan, S.L., & Bierer, L.M. (2001). Relationship 309311.
of parental trauma exposure and PTSD to PTSD, depressive
and anxiety disorders in offspring. Journal of Psychiatric Received: 8 July 2014
Research, 35 (5), 261270. Accepted: 14 October 2014

2014 John Wiley & Sons Ltd

Copyright of Developmental Science is the property of Wiley-Blackwell and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.